relating radiation-induced regional lung injury to changes in pulmonary function tests

PII S0360-3016(01)01619-4

CLINICAL INVESTIGATION Lung

RELATING RADIATION-INDUCED REGIONAL LUNG INJURY TO CHANGESIN PULMONARY FUNCTION TESTS

MING FAN, M.D.,*1 LAWRENCE B. MARKS, M.D.,* PEHR LIND, M.D. Ph.D.,* DONNA HOLLIS, M.S.,†

ROXANNE T. WOEL, B.A.,* GUNILLA G. BENTEL, R.N, R.T.T.,* MITCHELL S. ANSCHER, M.D.,*TIMOTHY D. SHAFMAN, M.D.,* R. EDWARD COLEMAN, M.D.,‡ RONALD J. JASZCZAK, M.D.,‡ AND

MICHAEL T. MUNLEY, PH.D.*

*Department of Radiation Oncology,†Cancer Center Biostatistics, and‡Department of Radiology, Duke University Medical Center,Durham, NC

Purpose: To determine whether the sum of radiotherapy (RT)-induced reductions in regional lung perfusion isquantitatively related to changes in global lung function as assessed by reductions in pulmonary function tests(PFTs).Methods and Materials: Two hundred seven patients (70% with lung cancer) who received incidental partial lungirradiation underwent PFTs (forced expiratory volume in 1 s and diffusion capacity for carbon monoxide) beforeand repeatedly after RT as part of a prospective clinical study. Regional lung function was serially assessedbefore and after RT by single photon emission computed tomography perfusion scans. Of these, 53 patients had105 post-RT evaluations of changes in both regional perfusion and PFTs, were without evidence of intrathoracicdisease recurrence that might influence regional perfusion and PFT findings, and were not taking steroids. Thesummation of the regional functional perfusion changes were compared with changes in PFTs using linearregression analysis.Results: Follow-up ranged from 3 to 86 months (median 19). Overall, a significant correlation was found betweenthe sum of changes in regional perfusion and the changes in the PFTs (p 5 0.002–0.24, depending on theparticular PFT index). However, the correlation coefficients were small (r 5 0.16–0.41).Conclusions: A statistically significant correlation was found between RT-induced changes in regional function(i.e., perfusion) and global function (i.e., PFTs). However, the correlation coefficients are low, making it difficultto relate changes in perfusion to changes in the PFT results. Thus, with our current techniques, the predictionof changes in perfusion alone does not appear to be sufficient to predict the changes in PFTs accurately.Additional studies to clarify the relationship between regional and global lung injury are needed. © 2001Elsevier Science Inc.

Radiation lung injury, Pulmonary function tests, Lung perfusion, Single photon emission computed tomography.

INTRODUCTION

The lung is one of the dose-limiting organs for patientsreceiving thoracic radiotherapy (RT). RT-induced pulmo-nary symptoms occur in approximately 5–20% of patientsirradiated for lung cancer, breast cancer, or Hodgkin’s dis-ease, and 50–90% experience declines in pulmonary func-tion tests (PFTs) (1, 2).

Because of the manner in which lung subunits are struc-

tured in parallel with each other, it is reasonable to assumethat changes in whole lung function (e.g., assessed by PFTs)would correlate with the sum of changes in regional lungfunction. We, and investigators at the Netherlands CancerInstitute (NKI), have used single photon emission computedtomography (SPECT) lung perfusion scans to relate changesin regional perfusion (assumed proportional to function) toregional dose. Independently, at both institutions, dose–response curves (DRCs) from many patients were summed

Reprint requests to: Lawrence B. Marks, M.D., Department ofRadiation Oncology, Box 3085, Duke University Medical Center,Durham, NC 27710. Tel: 919-668-5640; Fax: 919-684-3953; E-mail: [email protected]

Presented at the 42nd Annual meeting of the American Societyfor Therapeutic Radiology and Oncology, Boston, MA, October22–26, 2000.

Supported, in part, by PHS Grants R01-CA33541, R29-CA69579, and R01-CA76006, awarded by the National CancerInstitute, and Grant DE-FGO2-96ER62150, awarded by the De-partment of Energy.

Acknowledgments—Thanks to Phil Antoine, Robert Clough, andAndrea Tisch for computer and data management support, to JaneHoppenworth for secretarial support, and to the University ofNorth Carolina at Chapel Hill for the use of the PLUNC treatmentplanning system.1 Dr. Fan’s current address is Department of Radiation Oncology,Shanghai Cancer Hospital, Fu Dan University School of Medicine,Shanghai, China.

Received Jan 16, 2001, and in revised form Apr 2, 2001.Accepted for publication Apr 17, 2001.

Int. J. Radiation Oncology Biol. Phys., Vol. 51, No. 2, pp. 311–317, 2001Copyright © 2001 Elsevier Science Inc.Printed in the USA. All rights reserved

0360-3016/01/$–see front matter

311

to generate population-based DRCs (3, 4). Recent studiesfrom Duke University Medical Center (5) and NKI (6) haveattempted to quantitatively predict post-RT declines in PFTsby adding together predicted regional reductions in perfu-sion and/or ventilation based on the population DRC. Over-all, statistically significant correlations existed between thepredicted and measured reductions in PFTs (5, 6). However,the reported correlation coefficients (R) were only fair, sug-gesting that there could be limitations to this approach.

In this analysis, we related changes in the individualpatient’s SPECT scan (i.e., the individual DRC) withchanges in the PFTs. Thus, the present study differs fromthe prior (5) in which the population DRC was used topredict the sum of changes in regional perfusion that apatient would experience (Fig. 1). Because predictionsbased on the population DRC were only fair, we hereinexplore the possibility that patient-specific differences in thedose response of the lung could be responsible for animportant part of the variability noted in the prior analysis(5).

MATERIALS AND METHODS

Between 1991 and 1999, 207 patients were entered intoand followed on an Institutional Review Board–approvedprospective clinical study to assess the changes in regionaland whole lung function after thoracic RT at Duke Univer-sity Medical Center. Informed consent was obtained. Pa-tients underwent SPECT lung perfusion scans, CT and PFTsbefore and repeatedly (typically at 3–6-month intervals)after RT. Patients receiving high-dose chemotherapy, usingsteroids, or with intrathoracic recurrence that had an impacton lung function were excluded. Fifty-three remaining pa-tients with evaluations of changes in both regional perfusionand PFTs at approximately the same time after RT (0–4week time difference between 2 tests) were studied. The

follow-up range was 3–86 months (median 19). The patientcharacteristics are shown in Table 1.

General radiation techniqueThe radiation technique for patients enrolled in this pro-

tocol has previously been described (5). In brief, RT wasadministrated by 6–10 MV x-rays. Patients with lung can-cer were generally treated with opposed anterior and pos-terior fields to 40–45 Gy, followed by off-cord obliquefields to approximately 66 Gy, with 1.8–2.0 Gy daily frac-tions. Thirteen were treated using a hyperfractionated con-current boost technique: 1.25 Gy twice daily to the clinicaltarget volume and 1.6 Gy twice daily to the gross disease(minimum 6-h interfraction interval) to a total of 73.6 Gy.Patients with breast cancer were treated with tangentialfields to the chest wall or breast to 46–50 Gy, with 2-Gydaily fractions. Patients with lymphoma were treated withanterior and posterior opposed fields to approximately20–40 Gy, with 1.5–1.8 Gy daily fractions.

Pulmonary function testsThe PFTs included spirometry, volume measurements,

and diffusion capacity. This analysis was limited to theforced expiratory volume in 1 s (FEV1) and diffusion ca-pacity for carbon monoxide (DLCO) because these are mostoften cited (6–11). The latter was corrected for hemoglobinas follows unless otherwise noted: (corrected5 mea-sured3 [(10.221 hemoglobin)/(1.73 hemoglobin)].

For both indices (expressed as the percentage of thepredicted value, which was based on age, height, and sex),the measured follow-up data were compared to with thepre-RT values. The fractional loss in function [(12 post/pre) p 100] was the quantitative endpoint used.

Fig. 1. Calculation of the sum of regional reductions in perfusionbased on the DVH or DFH. The present study used patient-specificDRCs to calculate actual reductions in overall lung perfusion. Aprior study (5) considered the population DRC to predict thedegree of regional injury. The sum of regional injury (eitherpredicted or actually observed) was then compared with the mea-sured changes in whole lung function assessed by the FEV1,uncorrected DLCO, or DLCO.

Table 1. Patient characteristics*

Characteristic (n 5 53) Value

Sex (male/female) 25/28Median age (y) 60 (21–87)Tumor type

Lung cancer 36Breast cancer 8Lymphoma 6Other thoracic cancer 3

Median pack year tobacco 40 (0–125)Central tumor 26Pre-RT surgery 15Chemotherapy (before/concurrent with RT) 12/3Median percentage of predicted baseline PFTs

FEV1 69 (17–136)Uncorrected DLCO 68 (22–132)DLCO 75 (23–130)

Abbreviations:RT 5 radiotherapy; FEV1 5 forced expiratoryvolume in 1 s; DLCO5 diffusion capacity for carbon monoxide.

Numbers in parentheses are the range.* All values are nos. of patients unless otherwise noted.

312 I. J. Radiation Oncology● Biology ● Physics Volume 51, Number 2, 2001

Dose–volume histogram and dose–function histogramCT and SPECT perfusion images in the treatment posi-

tion were obtained before RT, as previously described (3,5). The 3-dimensional (3D) dose distributions were calcu-lated using tissue density inhomogeneity corrections. Thedose–volume histograms (DVHs) were calculated based onthe absolute total dose without adjustments for fraction sizeor overall treatment time.

The SPECT images were visually registered with theplanning CT and, hence, the dose distribution. The per-centage of SPECT counts in each dose bin (5-Gy inter-val) was used to generate a “dose SPECT-count histo-gram.” Assuming that perfusion was proportional tofunction, this histogram was termed the dose–functionhistogram (DFH) (3).

Furthermore, the pre-RT CT and SPECT images werevisually reviewed to assess the presence of hypoperfusionadjacent to a central (mediastinal or hilar) thoracic masscaused by either the primary tumor or an enlarged lymphnode (12). Each post-treatment SPECT scan was also

visually compared with the pre-RT SPECT image todetect reperfusion of previously underperfused areas thatmight be associated with restoration of lung function.These assessments were made independent of the PFTanalysis.

Calculation of DRC and reduction in perfusionEach patient’s pre- and post-RT DFHs were quantita-

tively compared to derive the patient-specific DRC at eachtime point, which demonstrated changes in regional perfu-sion vs. regional dose (3, 8). Assuming no RT-inducedchange in “low”-dose regions, the number of SPECT countsat all sites was normalized to the number of counts in thearea receiving#2.5 Gy (0-Gy dose bin). To assess whetherthe size of this normalization region had an impact on theresults, the data were recalculated using#7.5 (5-Gy dosebin) or#12.5 Gy (10-Gy dose bin) as the “low”-dose region(Fig. 2).

The percent change in perfusion throughout the lung was

Fig. 2. Calculation and normalization of a patient’s DRC. The 3D dose distribution was overlaid on the (A) pre- and (B)post-RT SPECT perfusion scans to generate the corresponding DFHs (C). Differences between the two DFHs were usedto calculate changes in the perfusion vs. regional dose (C, dashed area) to yield the unnormalized DRC for that patientat that time interval (D). Assuming no perfusion changes in a low-dose region (#2.5, 7.5, or 12.5 Gy), the DRC wasnormalized to zero in that area.

313Relating regional injury to changes in PFTs● M. FAN et al.

calculated by summing the actual reductions in regionallung perfusion:

Post-RT change in lung perfusion5 Od 5 0

dmax

(Vd * Rd)

where Vd is the volume of lung irradiated to dosed (fromthe DVH), and Rd is the reduction in regional perfusion atdosed (from the patient’s individual DRC). The calcula-tions were repeated using perfusion-weighted volumes atdosed (PVd) derived from the DFH. Because the sums ofreduction in regional perfusion calculated on the basis of theDVH or DFH correlate highly with each other (r 5 0.97),only the results based on the DVH are presented.

Statistical analysisThe calculated change in regional perfusion was com-

pared with the reductions in PFTs (Fig. 1), using both the

Pearson product-moment correlation and linear regressionanalysis.

RESULTS

Fifty-three patients had a total of 105 concurrent post-RTevaluations of changes in both SPECT perfusion scans andPFTs. The comparisons of the sum of reductions in regionallung perfusion and the reductions in PFTs for all matchesare shown in Fig. 3. Significant correlations (p , 0.001–0.04) were revealed between reductions in perfusion andPFTs.

In 33 patients,.1 post-RT measurement was performed.Multisampling in a single patient may introduce bias to theresults. Thus, the data were reanalyzed, and multiple datapoints from a single patient were averaged. The compari-sons of the averaged sum of reductions in regional lungperfusion and the averaged reduction in PFTs for the overallgroup are shown in Fig. 4. The correlation between the sum

Fig. 3. (A–C) Comparison between the DVH-based perfusion reductions and percent reduction in PFTs for allobservations (n 5 105). The reference line ofy 5 x is included.

Fig. 4. (A–C) Comparison between DVH-based perfusion reductions and percent reduction in PFTs for all patients (n 553); however, the results for the 33 patients with.1 observations were averaged.


of perfusion reductions and the reduction in FEV1 wassignificant (p 5 0.002). The correlation was not significantfor uncorrected DLCO or DLCO (p 5 0.24 and 0.22,respectively). Furthermore, the correlation coefficients weremodest for all three comparisons (r 5 0.16–0.41).

Restricting the analysis to patient subgroups according topre-RT PFT level (over average vs. less than average),chemotherapy exposure, tumor type (lung vs. non–lung), orthe presence or absence of a central tumor and adjacenthypoperfusion did not appear to markedly influence theresults. Furthermore, the results did not change if the DRCwas normalized to the 5 or 10-Gy dose bins.

DISCUSSION

RT-induced pulmonary injury is common after treatmentfor tumors in and around the thorax. Currently, no meansare available to predict RT-induced changes in PFTs accu-rately.

In patients having portions of their lung resected, severalstudies haved related changes in PFTs to the percentage ofperfused lung resected (13–18) (Table 2). Cordineret al.(13) reviewed 18 patients undergoing lobectomy or pneu-monectomy for lung cancer. They noted a strong correlation(p ,0.001,r 5 0.82) between the approximate percentageof perfused lung resected (assessed by planar lung perfusionscans) and the percentage of reduction in FEV1. Similarly,Pierce et al. (14) studied 45 patients with lung cancer,whose postoperative FEV1 and DLCO correlated highlywith their remaining functioning lung (r 5 0.87 and 0.56respectively;p ,0.0001). Giordaioet al. (15) found asimilar result in a group of 41 patients who underwentlobectomy (r 5 0.87 for FEV1). In concert, these studiessuggest that quantification of the percentage of perfusionascribed to a region of the lung can be related to thepercentage of loss in whole lung function (as assessed byPFTs) that the patient would experience after resection. Thisapproach is particularly reasonable in patients undergoingresection, because resected regions of the lung contribute nofunction after resection, and the unresected regions retainfull function. The surgical insult to the lung, therefore, canbe considered as a step function, all or none.

In the 1980s and early 1990s, investigators at Massachu-setts General Hospital (10, 19) and the Fox Chase CancerCenter (20, 21) tried to correlate RT-induced reductions inFEV1 with the estimated percentage of perfused lung irra-diated. In the MGH study, the “proportion of the function-ing lung that would be damaged” was estimated by review-ing the simulation films, planar perfusion/ventilation scans,and$40–45 Gy isodose line on the dose distribution (10,19). In the Fox Chase study (21), the percent of perfusedlung within the AP/PA field (generally receiving 40–45 Gy)was used as a surrogate for the “fraction of functioning lung”irradiated. In both of these studies, the results were mixed. Inthe study by Choiet al., the correlation rate varied from 10%to 70%, depending on the FEV1 reserve and the ventilation/perfusion shift to the uninvolved lung (10, 19). Curranet

Tab

le2.

Sel

ecte

dse

ries

rela

ting

estim

ated

perc

enta

geof

func

tiona

llun

gre

sect

edw

ithsu

rger

y-in

duce

dre

duct

ion

inP

FT

s

Aut

hor

Eva

luat

edP

atie

nts*

(n)

Sur

gica

lPro

cedu

reE

quat

ion

for

pred

icte

dpo

stop

erat

ive

PF

Ts

†

Pos

tope

rativ

eP

FT

sre

peat

ed(m

o)

FE

V1

DLC

O

rp

rp

Juliu

se

ta

l.(1

6)9

Pne

umon

ecto

my

A1–

21(m

ean

8.5)

0.80

,0.

02C

ordi

nere

ta

l.(1

3)18

Pne

umon

ecto

my

3A

3–4

0.82

,0.

001

Lobe

ctom

y15

Pie

rcee

ta

l.(1

4)45

Pne

umon

ecto

my

7A

(Pre

oper

ativ

eF

EV

1,

55%

)4

0.87

,0.

0001

0.56

,0.

0001

Lobe

ctom

y27

B(P

reop

erat

ive

FE

V1

.55

%)

Wed

geR

esec

tion

11B

ollig

ere

ta

l.(1

7)22

Pne

umon

ecto

my

5A

3,6

0.81

(3m

o)0.

74(3

mo)

Lobe

ctom

y15

Wed

geR

esec

tion

20.

84(6

mo)

0.74

(6m

o)G

iord

aio

et

al.

(15)

41Lo

bect

omy

C1

0.87

Ab

bre

via

tion

s:P

FT

s5pu

lmon

ary

func

tion

test

s;P

PO5

pred

icte

dpo

stop

erat

ive;

othe

rab

brev

iatio

nsas

inT

able

1.

*P

atie

nts

that

died

beca

use

ofco

mpl

icat

ions

wer

eex

clud

edfr

omth

eor

igin

alre

port

,be

caus

eth

eir

post

oper

ativ

eP

FT

sw

ere

not

avai

labl

e.†

A)

pred

icte

dpo

stop

erat

ive

PF

T5pr

eope

rativ

eP

FT3

%pe

rfus

ion

ofre

mai

ning

lung

;B

)P

PO

PF

T5pr

eope

rativ

eP

FT3

(12

#of

segm

ents

rem

oved4

19);

C)

PP

OP

FT5

preo

pera

tive

PF

T3(1

2#

ofse

gmen

tsre

mov

ed4#

ofse

gmen

tsin

that

lung

[10

“fun

ctio

nal”

(not

anat

omic

)se

gmen

tsin

the

right

lung

,9

inth

ele

ftlu

ng,

asde

fined

byM

iorn

er(1

8)])

.


al. (Fox Chase) found their technique underestimated thepost-RT FEV1 in approximately 90% of patients (r or p valuenot provided). In a prospective study, Abrattet al. (22) re-ported no statistically significant correlation between de-creased perfusion and the reduction in FEV1, DLCO, or forcedvital capacity at 6 and 12 months post-RT in a group of 36patients with lung cancer. These studies did not have thebenefit of 3D treatment planning. The “percentage of perfusedlung irradiated” was estimated in each of the studies in aslightly different manner. The approach taken was analogousto that used by the surgeons (i.e., a step function). Lungirradiated to “most of the dose” was assumed to lose all of itsfunction, and the remainder of the lung was assumed to retainits function. In reality, however, the radiation delivered washeterogeneous, and probably resulted in different degrees ofregional lung function within the radiation field.

Using modern 3D planning, we, and investigators at theNKI (3, 23), have demonstrated dose-dependent changes inregional lung function. Rather than observing a step func-tion, both institutions have noted a gradual reduction in lungperfusion with in the range of 15–60 Gy.

With the improved knowledge relating changes in re-gional perfusion to regional radiation dose, we have hypoth-esized that one should be able to relate the changes in wholelung function to the sum of changes in regional perfusion. Ina prior study (5), we related the sum of predicted changes inregional lung perfusion with the maximal decline in PFToccurring after RT. In that analysis, each individual pa-tient’s DFH was “multiplied” by the population-based DRCfor radiation-induced, regional lung dysfunction (Fig. 1).Although a statistically significant correlation was foundbetween the predicted and actual decline in PFT (p 50.005–0.08), a lot of scatter was noted in the data (r 50.18–0.30). The investigators at NKI have reported resultsfrom a similar approach in a group of 81 patients withlymphoma or breast cancer (6). They obtained a bettercorrelation (p ,0.001,r 5 0.57–0.75).

We previously reported patient-specific differences in theDRC for radiation-induced changes in regional perfusion (3).Although we could not specifically ascribe differences in thepopulation-DRC to patient-specific factors (e.g., tobacco use,chemotherapy exposure, baseline PFTs), interpatient differ-ences were still noted, which could have been due to differ-ences in radiation sensitivity. In this study, we considered thepossibility that the inability of our prior study to relate changesin PFTs to the sum of predicted change in regional functionwas because we used the population-based DRC, rather thanthe individual DRC. Therefore, in this study, we used eachpatient’s individual DRC to determine that patient’s sum ofactual changes in regional perfusion. With this approach, eachpatient could be considered at different time points, corre-sponding to each post-RT evaluation.

Our results demonstrated that the correlation between thepatient’s actual sum of reductions in regional perfusion andtheir change in PFT was not better than the prediction basedon the population-based DRC. Therefore, it appears that

techniques that accurately predict changes in regional per-fusion may not be adequate in predicting changes in PFTs.

Furthermore, these data suggest that predictions based onthe population DRC are better predictors than the patient-specific DRC. This result is probably because the calcula-tion of the individual patient’s DRC is sometimes inaccuratebecause of patient motion, penumbra, divergence effects,and registration. The development of the population-basedDRC, in which each patient’s contribution to a particulardose level is proportional to the percentage of their lungirradiated at that dose, is an attempt to address the inaccu-racies of the individual patient’s DRC. The population-based DRCs may therefore be more accurate than the indi-vidual patient DRCs. The population-based DRC derived atDuke University Medical Center from 50 patients is essen-tially equivalent to that derived at the NKI (4).

It remains difficult to relate the sum of regional perfusionchanges (either actual or predicted) with measured changesin PFTs. The reason for this is not clear, but the possibilitiesare several. First, after thoracic irradiation, the function insome regions of the lung may actually improve because oftumor shrinkage (10, 12) and/or retraction of irradiatedlung, enabling expansion of adjacent healthy regions of lung(similar to that achieved with reduction surgery for emphy-sema). Second, regional blood flow might not represent thebest estimate of regional lung function. Gas exchange re-quires both adequate aeration and adequate perfusion, withan intact alveolar capillary unit for effective gas diffusion.Our studies did not consider changes in ventilation. Weelected to focus on perfusion because the available literaturesuggests perfusion to be a more sensitive marker of radia-tion-induced injury (10, 19). That is, in areas of the lung inwhich ventilation is reduced, the normal physiologic re-sponse of the area capillaries is to constrict, causing reducedperfusion, and thus reducing mismatch (24).

Third, it is possible that PFTs measured in the manner usedin the study are not ideal assessments of whole lung function.The presence of symptoms does not always track well withchanges in PFTs. In our series of patients, shortness of breathoccurred in only 6 of 15 (22%) of 27 patients with a.20%decline in FEV1. On the other hand, 13 (87%) of 15 patientswith RT-induced shortness of breath experienced a decline inPFT (2). Abrattet al. (25) noted a worsening dyspnea scoreonly in patients who had a.10% decline in DLCO. In a seriesof 1277 smokers with COPD (91% with any pulmonary symp-tom), only 36% complained of dyspnea on exertion (26).Furthermore, shortness of breath is a subjective endpoint, andpatients with chronic disease are less likely to complain thanare those with an acute illness (27). Alternative assessments ofwhole lung function might be more functionally relevant to thepatient, such as exercise tolerance. We are considering modi-fying our protocols to include these more physiologic/func-tional measurements.

It is also possible that the mechanism of RT-induced lunginjury differs over time after radiation. In the early phases,a radiation-induced reduction in function might be mostrelated to damage to type II pneumocyte and endothelial


cells of the microvasculature, leading to leaky vessels, ex-udate, and decreased permeability of the alveolar spaces(28). Later, more of a fibrotic reaction occurs, leading todisruption of the normal architecture of the alveolar capil-lary network with resultant loss in gas exchange. Interest-ingly, in this study, when we look at the 105 observationsdivided according to the interval after RT, a slightly bettercorrelation coefficient was seen for DLCO at#3 monthscompared to longer intervals (r 5 0.37 vs. 0.18).

In our analysis, the sum of actual changes in regional per-fusion correlated better with changes in FEV1 than to changesin DLCO. Interestingly, the surgical series (Table 2) and NKIstudy (6) also noted betterr values for the prediction of FEV1

than for DLCO. Conversely, in our prior analysis (5), the sumof predicted regional injuries was correlated better withchanges in DLCO than FEV1. The endpoint in our prior studywas the maximal decline in post-RT PFT.

The results of this study were based on only 53 evaluatedpatients. One of the difficulties in conducting research in pa-tients with lung cancer is the mortality/morbidity of the dis-ease, reducing the number of long-term observations. Manypatients are lost because of death from recurrent cancer or arenot evaluated because recurrent cancer has an impact on pul-monary function. Additional studies involving a larger numberof patients are needed to better understand the relationshipbetween the changes in regional and whole lung function.

REFERENCES

1. Marks LB. Pulmonary effects of thoracic irradiation.Oncol-ogy 1994;8:89–100.

2. Marks LB, Fan M, Clough R,et al. Radiation induced pul-monary injury: Symptomatic versus subclinical endpoints.IntJ Radiat Biol Phys2000;76:469–475.

3. Garipagaoglu M, Munley MT, Hollis D,et al. The effect ofpatient-specific factors on radiation-induced regional lung in-jury. Int J Radiat Oncol Biol Phys1999;45:331–338.

4. Theuws JCM, Kwa SLS, Wagenaar AC,et al. Dose-effectrelations for early local pulmonary injury after irradiation formalignant lymphoma and breast cancer.Radiother Oncol1998;48:33–43.

5. Fan M, Marks LB, Hollis D,et al. Can we predict radiation-induced changes in pulmonary function tests based on the sum ofpredicted regional dysfunction?J Clin Oncol2001;19:543–550.

6. Theuws JCM, Kwa SLS, Wagenaar AC,et al. Prediction ofoverall pulmonary function loss in relation to the 3-D dosedistribution, for patients with breast cancer and malignantlymphoma.Radiother Oncol1999;49:233–243.

7. Mah K, Keane TJ, Van Dyk J,et al. Quantitative effect ofcombined chemotherapy and fractionated radiotherapy on theincidence of radiation-induced lung damage: A prospectiveclinical study.Int J Radiat Oncol Biol Phys1994;28:563–574.

8. Marks LB, Munley MT, Bentel GC,et al. Physical and bio-logical predictors of changes in whole lung function followingthoracic irradiation.Int J Radiat Oncol Biol Phys1997;39:563–570.

9. Brady LW, Germon PA, Cander L. The effects of radiationtherapy on pulmonary function in carcinoma of the lung.Radiology1965;85:130–134.

10. Choi NC, Kanarek DJ. Toxicity of thoracic radiotherapy onpulmonary function in lung cancer.Lung Cancer 1994;10(Suppl. 1):S219–S230.

11. Choi NC, Kanarek DJ, Grillo HC. Effect of postoperativeradiotherapy on changes in pulmonary function in patientswith stage II and IIIA lung carcinoma.Int J Radiat Oncol BiolPhys1990;18:95–99.

12. Marks LB, Hollis D, Munley MT,et al. The role of lungperfusion imaging in predicting the direction of radiation-induced changes in pulmonary function tests.Cancer2000;88:2135–2141.

13. Cordiner A, de Carlo F, de Gennaro R,et al. Prediction ofpostoperative pulmonary function following thoracic surgeryfor bronchial carcinoma.Angiology1991;42:985–989.

14. Pierce RJ, Copland JM, Sharpe K,et al. Preoperative riskevaluation for lung cancer resection: Predicted postoperativeproduct as a predictor of surgical mortality.Am J Respir CritCare Med1994;150:947–955.

15. Giordano A, Calcagni ML, Meduri G,et al. Perfusion lung

scintigraphy for the prediction of postlobectomy residual pul-monary function.Chest1997;111:1542–1547.

16. Julius AJ, de Jong D, van Deutekom H,et al.The value of99mTcmacroaggregated albumin lung perfusion scanning in the predic-tion of postpneumonectomy function and pulmonary artery pres-sure.Scand J Thorac Cardiovasc Surg1987;21(1):81–85.

17. Bolliger CT, Wyser C, Roser H,et al. Lung scanning andexercise testing for the prediction of postoperative perfor-mance in lung resection candidates at increased risk for com-plications.Chest1995;108:341–348.

18. Miomer G. 133Xe-radiospirometry: A clinical method forstudying regional lung function.Scand J Respir Dis1968;(Suppl. 64):1–84.

19. Choi NC, Kanarek DJ, Kazemi H. Physiologic changes inpulmonary function after thoracic radiotherapy for patientswith lung cancer and role of regional pulmonary functionstudies in predicting post-radiotherapy pulmonary functionbefore radiotherapy.Cancer Treat Sympos1985;2:119–130.

20. Rubenstein JH, Richter MP, Moldofsky PJ,et al. Prospectiveprediction of post-radiation therapy lung function using quan-titative lung scans and pulmonary function testing.Int J Ra-diat Oncol Biol Phys1988;15:83–87.

21. Curran WJ Jr, Moldofsky PJ, Solin LJ. Observations on thepredictive value of perfusion lung scans on post-irradiationpulmonary function among 210 patients with bronchogeniccarcinoma.Int J Radiat Oncol Biol Phys1992;24:31–36.

22. Abratt RP, Willcox PA. Changes in lung function and perfu-sion after irradiation in patients with lung cancer.Lung Can-cer 1994;11:61–69.

23. Seppenwoolde Y, Muller SH, Theuws JC,et al. Radiationdose-effect relations and local recovery in perfusion for pa-tients with non-small-cell lung cancer.Int J Radiat Oncol BiolPhys2000;47:681–690.

24. Comroe JH: The pulmonary circulation. In: Comroe JH, ed.Physiology of respiration. Chicago: Year Book Medical Pub-lishers; 1965.p.129–135.

25. Abratt RP, Willcox PA, Smith JA. Lung cancer in patientswith borderline lung functions—zonal lung perfusion scans atpresentation and lung function after high dose irradiation.Radiother Oncol1990;19:317–322.

26. Calverley PM. COPD: Early detection and intervention.Chest2000:117(Suppl. 2):365S–371S.

27. Burdon JG, Pain MC, Rubinfeld AR,et al. Chronic lungdiseases and the perception of breathlessness: A clinical per-spective.Eur Respir J1994;7:1342–1349.

28. McDonald S, Rubin P, Phillips TH,et al. Injury to the lungfrom cancer therapy: Clinical syndromes, measurable end-points, and potential scoring systems.Int J Radiat Oncol BiolPhys1995;31:1187–1203.


relating radiation-induced regional lung injury to changes in pulmonary function tests

Documents