High resolution computerized tomography of the chest and pulmonary function testing in evaluating the effect of tobramycin solution for inhalation in cystic fibrosis patients.

Abstract

To evaluate the sensitivity of high-resolution computerized tomography (HRCT) of the chest compared to spirometry measures in evaluating the effects of tobramycin solution for inhalation (TSI) in cystic fibrosis (CF) patients.Thirty-two subjects >/=6 years old with mild to moderate CF lung disease were enrolled in a randomized, double-blind, placebo-controlled pilot study. Duration was 28 days; 31 subjects completed the study.HRCT scores decreased 4.06 +/- 3.20 (mean +/- SD) for TSI and decreased 0.17 +/- 1.78 for placebo subjects (P = 0.13). Mean forced expiratory flow during middle half of forced vital capacity (FEF(25%-75%)) predicted increased 6.08 +/- 4.86 for TSI and decreased 0.60 +/- 2.34 for placebo (P = 0.23). Percentage forced expiratory volume in 1 s (FEV(1)) predicted increased slightly for both TSI and placebo (1.29 +/- 3.33 for TSI and 1.17 +/- 1.4 for placebo) (P = 0.97). Two of eight HRCT component scores (atelectasis and inhomogeneity) were observed to be highly discordant with observed HRCT global total score and other HRCT component scores. A modified total score was calculated by dropping them from the global total score. The modified HRCT total scores decreased 6.68 +/- 3.09 for TSI subjects and increased 0.02 +/- 2.0 for the placebo subjects (P = 0.07). Sample sizes were calculated to show statistical significance by differences in modified total HRCT scores, global total HRCT scores, FEF(25%-75%) predicted or FEV(1) % predicted. A total of 60, 100, 200, and over 800 patients would be necessary respectively.HRCT can be a useful measure of change in CF pulmonary disease, requiring a smaller sample size than that required to show treatment effect by pulmonary function testing (PFT) alone.

Full text

Page 1

Pediatric Pulmonology 41:1129–1137 (2006)
High Resolution Computerized Tomography of the Chest
and Pulmonary Function Testing in Evaluating the
Effect of Tobramycin Solution for Inhalation
in Cystic Fibrosis Patients
Samya Z. Nasr, MD,1* Diane Gordon, RN, MS, MPH,4Ermelinda Sakmar, MA, RN,1Xin Yu, PhD,5
Emmanuel Christodoulou, PhD,3Boris P. Eckhardt, MD,2and Peter J. Strouse, MD2
Summary. Toevaluatethesensitivityofhigh-resolutioncomputerizedtomography(HRCT)ofthe
chest compared to spirometry measures in evaluating the effects of tobramycin solution for
inhalation (TSI) in cystic fibrosis (CF) patients.Thirty-two subjects ?6 years old with mild to
moderate CF lung disease were enrolled in a randomized, double-blind, placebo-controlled pilot
study.Durationwas28days;31subjectscompletedthestudy.HRCTscoresdecreased4.06?3.20
(mean?SD) for TSI and decreased 0.17?1.78 for placebo subjects (P¼0.13). Mean forced
expiratory flow during middle half of forced vital capacity (FEF25%–75%) predicted increased
6.08?4.86 for TSI and decreased 0.60?2.34 for placebo (P¼0.23). Percentage forced
expiratory volume in 1 s (FEV1) predicted increased slightly for both TSI and placebo
(1.29?3.33 for TSI and 1.17?1.4 for placebo) (P¼0.97). Two of eight HRCTcomponent scores
(atelectasisandinhomogeneity)wereobservedtobehighlydiscordantwithobservedHRCTglobal
total score and other HRCTcomponent scores. A modified total score was calculated by dropping
them from the global total score. The modified HRCT total scores decreased 6.68?3.09 for TSI
subjects and increased 0.02?2.0 for the placebo subjects (P¼0.07). Sample sizes were
calculatedtoshowstatisticalsignificancebydifferencesinmodifiedtotalHRCTscores,globaltotal
HRCTscores, FEF25%–75%predicted or FEV1% predicted. A total of 60, 100, 200, and over 800
patients would be necessary respectively.HRCT can be a useful measure of change in CF
pulmonarydisease, requiringa smaller sample size thanthat required toshow treatment effectby
pulmonary function testing (PFT) alone. Pediatr Pulmonol. 2006; 41:1129–1137.
? 2006 Wiley-Liss, Inc.
Key words: cystic fibrosis; HRCT; FEV1; FEF25%–75%; TSI; Pseudomonas aeruginosa.
INTRODUCTION
Progressive destruction of the lungs leads to increased
morbidity and mortality in patients with cystic fibrosis
(CF). Inflammation and chronic infection of the small
airways lead to pulmonary dysfunction.1High resolution
computed tomography of the chest (HRCT) can illustrate
the airway involvement characteristic of this disease,
especiallyearlychanges,andthereforehasthepotentialto
beausefuloutcomemeasureinCF.2Itisnoninvasive,does
not require contrast material, and with limited number of
cuts, the overall radiation dose is low.3HRCTappears to
1Department of Pediatrics, Division of Pediatric Pulmonology, University
of Michigan Health System, Ann Arbor, Michigan.
2Department of Radiology, Section of Pediatric Radiology, University of
Michigan Health System, Ann Arbor, Michigan.
3Section of Physics/Engineering, University of Michigan Health System,
Ann Arbor, Michigan.
4Chiron Corporation. Currently at Department of Biobehavioral Nursing
and Health Systems, School of Nursing, University of Washington, Seattle,
Washington.
5Chiron Corporation, Seattle, Washington.
Presented in part at the European Respiratory Society Annual Congress,
Glasgow, Scotland from September 4 to 8, 2004, and at North American
Cystic Fibrosis Conference, St. Louis, MO from October 14 to 18, 2004.
Grant sponsor: Chiron Corp.
*Correspondence to: Samya Z. Nasr, MD, Associate Professor of
Pediatrics, 1500 E. Medical Center Drive, Ann Arbor, MI 48109-0212.
E-mail: snasr@umich.edu
Received 18 October 2005; Revised 7 March 2006; Accepted 7 March
2006.
DOI 10.1002/ppul.20447
Published online in Wiley InterScience
(www.interscience.wiley.com).
? 2006 Wiley-Liss, Inc.

Page 2

be a more sensitive measure than CXR in evaluating CF
lung disease, especially in patients with mild lung
disease.4
A recent study demonstrated positive correlation
between five different HRCT scoring systems and
pulmonary function testing (PFTs), especially forced
expiratory volume in 1 s (FEV1), forced expiratory flow
during middle half offorced vital capacity (FEF25%–75%),
FEV1/FVC ratio, and TLC values.5Correlation of the
scoring systems with each other and with PFT parameters
was very good.5Currently, PFTs are the most accepted
proven outcome measure for the evaluation of CF lung
disease. However, because of the slow decline in PFTs,
large patient samples are required in clinical trials to
detect significant changes when using these measures.6,7
Tobramycin solution for inhalation (TSI) has proven
effective in CF patients with chronic Pseudomonas
aeruginosa (P. aeruginosa) endobronchial infections.8
In CF patients with mild lung disease and chronic
P. aeruginosa lung infection, TSI leads to a significant
reduction in respiratory hospitalizations, concomitant
antibiotic use, and a trend towards improvement in
percent predicted forced expiratory flow (FEF25%–75%).9
The aim of this pilot study was to assess the sensitivity
of HRCTof the chest compared to PFT in evaluating the
effects of TSI in CF patients 6 years of age and older with
mostly mild to moderate lung disease.
MATERIALS AND METHODS
Study Design
A randomized, double-blind, placebo-controlled pilot
study was conducted. The study procedures and measures
were reviewed and approved by the University of
Michigan Institutional Review Board and consent was
obtained from patients and/or parents/guardians prior to
enrollment in the study.
Study Population
Patients with CF 6 years of age and older with FEV1of
45% predicted or higher, attending the University of
Michigan CF Center, were considered eligible for the
study. Thirty-two patients (12 males and 20 females) with
a confirmed diagnosis of CF by sweat chloride test and/or
genotype, and P. aeruginosa colonization in sputum
within 6 months of screening and at enrollment were
recruited. Exclusion criteria included known hypersensi-
tivity to aminoglycosides or history of Burkholderia
cepacia isolated from respiratory cultures. Patients had to
be clinically stable prior to enrollment as defined by last
hospitalization for pulmonary exacerbation and intrave-
nous antibiotic use at least 2 months prior to enrollment.
Oralantibioticscouldnotbeusedforatleast1monthprior
to enrollment; TSI could not be used for at least 3 months
prior to enrollment. All subjects were pancreatic insuffi-
cient.
Study Procedures
Study duration was 28 days and patients were enrolled
between February 2002 and August 2003. Patients were
evaluated at baseline and again at the end of the study.
Eligiblesubjectswererandomizedtoreceiveeither5mlof
TSI (treatment) or 5 ml of placebo twice a day, via
inhalation therapy. TSI (Chiron Corporation, Emeryville,
CA) is a solution of 300 mg tobramycin and 11.25 mg
sodium chloride in sterile water; the placebo was
chosen for its similar taste to tobramycin and consisted
of 1.25 mg quinine sulfate in normal saline. Standard
routine CF care was provided for both groups as
directed by the primary pulmonologist. Study medication
and placebowere administeredusinga Pari LCPlusTMjet
nebulizer and a Pulmo-Aide1compressor. Both TSI and
placebo were supplied by Chiron Corporation. The
University of Michigan investigational drug services
performed the double blinding for the study.
HRCT and spirometry (Sensormedics, Vmax 6200,
Yorba Linda, CA), were obtained at baseline and at the
end of the study. Spirometry was obtained according to
ATS guidelines. CXR was done at baseline for evalua-
tion of lung disease and as part of routine care. Throat/
sputum cultures were obtained prior to enrollment for
confirmation of colonization with P. aeruginosa. A
quality of life questionnaire (cystic fibrosis question-
naire (CFQ)), validated for use in CF patients,10was
administered at baseline and at the end of the study to
patients and parents/guardians. Adverse events, hospi-
talizations, concomitant medication use, and school
days/work days missed due to illness, were collected at
the end of the study.
HRCT scan was performed using 40 mA, 1 s gantry
rotation time, and 120 KVp for each slice through the
lung.UtilizingaHiSpeedAdvantage1scanner(General
Electric Medical Systems, Milwaukee, WI), images
were obtained at 1 mm thickness. Five equally spaced
images in inspiration and in expiration, after an AP
inspiratory scout view, were obtained for a total of
ABBREVIATIONS
CF
FEV1
FEF25%–75%
Cystic fibrosis
Forced expiratory volume in 1 s
Forced expiratory flow during middle half of forced
vital capacity
High resolution chest tomography
Hounsfield unit
Pulmonary function tests
P. aeruginosa Pseudomonas aeruginosa
TSI Tobramycin solution for inhalation
CFQCystic fibrosis questionnaire
HRCT
HU
PFTs
1130 Nasr et al.

Page 3

10 images. Patients were placed in the supine position
and instructed to take a deep breath and hold it for 7–
10 s. During each breath hold, the five inspiratory
sections were obtained. The same technique was
repeated after patients were asked to exhale to lowest
volume and hold their breath for the expiratory sections.
Patients practiced the techniques prior to the testing and
weremonitoredbyboththeradiologytechnicianandthe
research coordinator. If techniques were suboptimal,
patients were instructed to repeat them. The superior-
most image was at the upper margin of the posterior
3rd rib and the inferior-most image was 1 cm above the
diaphragm.Therangeoftheeffectiveradiationdosewas
estimated to be 0.03–0.05 mSv and the equivalent
radiation dose to the entire lung was estimated to be
0.09 mGy. HRCT images were reviewed on a work-
station (Sienet Magicview, version 42, Siemens, New
York, NY). Images were initially displayed with a
window width of 1,000 Hounsfield Units (HU) and level
of ?600 HU; however, the readers were permitted to
modify the window and level of display of individual
images and examinations for optimal viewing.
ScoringoftheHRCTscansandCXRswereperformed
by two pediatric radiologists (PJS and BPE). Both
readerswereblindedtopatientname,number,andstudy
date at the time of reading and scoring. Each CXR
(conventional analog) was scored twice by each radi-
ologist using the Brasfield scoring method,11and the
scores were averaged across the two readers. A CXR
score of 25 points indicates a normal finding, the score
decreasing with increasing severity of lung disease. The
HRCT score was derived from Brody, A.S. and
others.7,12–15Each reader read each HRCT twice in a
randomized, blinded fashion. Means of the scores for
each reviewer were calculated separately and compared
between baseline and end of study. Scores of both
readerswerealsocombined(fourtotalreadings);amean
of the combined score was calculated and compared
betweenbaselineandatendofstudy.Onereaderreadthe
HRCT scans for the second timewithin 1–2 weeks from
thefirstreading.ThesecondreaderreadtheHRCTscans
for the second time 1–2 months after the first reading.
Eachlung(right &left)was scoredforeachofthefive
imaging levels. Eight imaging component categories
werescoredat eachimaginglevel foreachlung: nodular
opacities, linear opacities, hyperinflation (lack of
normal decrease in volume on expiratory views),
peribronchial thickening (abnormal thickening of bron-
chial walls), atelectasis (areas of collapsed, non-aerated
lung), inhomogeneity (variation in aeration producing
variation in lung opacity), bronchiectasis (dilated
bronchi usually judged relative to diameter of adjacent
pulmonary artery branch), and mucus plugging (opaci-
fied bronchus/bronchi)... Hyperinflation and inhomo-
geneity were scored on expiration because the
expiratory images showed this finding better. Some
inspiratory images showed inhomogeneity as well. The
other six component categories were scored on inspira-
toryimages.Scoringrangedfrom0to4with0indicating
a normal finding and 4 indicating severe disease. The
HRCT final score of each patient was calculated
based on the severity and extent of the morphologic
changes, and represented the average score from
the two lungs and two readers. Three summary scores
were calculated. (1) A component category score: the
average of both reader’s scores across the five levels
within each category and for both lungs, (2) An imaging
level score: The average of both reader’s scores across
the eight components within each level and for both
lungs, (3) total global score: the sum of all eight
total HRCT category scores or the sum of all five level
scores.
Statistical Analysis
This is a pilot study. Power calculations of sample size
were not done. The number of subjects studied was
arbitrarily chosen. Frequencies and percentages were
calculated for categorical variables; mean, standard
deviation, minimum, and maximum were calculated for
continuous variables. The baseline analysis was based on
16patientsineachtreatmentgroup.Endofstudyanalyses
included data from 16 patients in the TSI arm and
15 patients in the placebo arm (excluding the drop-out
subject in the placebo arm). SAS version 8 software was
utilized for all analyses. A two-sided student’s t-test for
twoindependentsampleswasusedtoassessthetreatment
effect on the primary efficacy variable at a 0.05
significance level. In addition, non-parametric test
(Wilcoxon) was performed. Three summary scores were
calculated for the HRCT: a category score, a level score,
andatotalglobalscore.Thisscoringsystemwasbasedon
that used previously.7,12Additionally, maximal score was
determinedforbothcategoryandlevelscores.Coefficient
ofvariationandtestofdifferencesinmeanscoresbetween
and within readers (mean?SD) were used to measure
inter/intra-reader variability in scoring CXR and HRCT
images. Inter/intra-rater reliability analyses were carried
out to test whether there was a difference between the
reliabilityofthescoringforHRCTandCXR.Thesubjects
in both groups were combined to enlarge the sample size
for statistical analyses. HRCT scores from both visits and
CXR scores from the first visit were averaged for these
analyses.Coefficientofvariationandtestofdifferencesin
mean scores between and within readers (mean?SD)
were done for the total HRCT score, all HRCT subscores,
and the CXR scores.
Correlation statistics (R2) were used to measure
correlation between specific HRCT component scores
and the global total HRCT scores. The coefficient of
HRCT in Evaluating TSI in Cystic Fibrosis 1131

Page 4

variation was determined according to the following
formula:
Absolutemeandifference
¼?100ðexpressedasapercentageÞ
Overallmeanof scores
where the ‘‘absolute mean difference is the mean of the
absolute value in the differences in the scores of the two
readers, and the ‘‘overall mean of scores’’ is the mean of
the scores of the two readers for all subjects included. A
frequency distribution of HRCT score differences was
alsoconductedforthe80individualscores,bypatientand
by case (there were 5,009 cases, over two visits that were
evaluable)todemonstratethepatternofscoringforeachof
the two readers. Ninety-seven percent of the observations
had a maximum score difference of 1 or less. Maximum
difference¼(maximum of the four scores, E1, E2, S1,
S2)?(minimum of the four scores E1, E2, S1, S2). Only
nine of the observations had maximum score differences
of 3 or more. Correlations between specific HRCT
changes and total HRCT scores: Correlation statistics
measure the degree of linear relationship between two
variables. Pearson Product Moment Correlation Coeffi-
cient is used for continuous variables. An R-value closer
to1indicatesastrongpositivecorrelation;avaluecloseto
?1indicatesastrongnegativecorrelation.Valuescloseto
zero indicate there is no linear relationship between the
two data sets.
RESULTS
Thirty-two patients were enrolled and randomized to
one of the two treatment groups. Thirty-one patients
completed the study. One patient in the placebo group
droppedoutofthestudyafterfewdayscomplainingabout
thetasteofthe‘‘medication.’’Patientswereaskedtobring
empty vials at the end of the study to evaluate adherence
rate. Adherence rate was calculated at 86% for the study
patients.
Baseline clinical characteristics of the two groups are
presented in Table 1. Most of the study population had
mild to moderate lung disease as indicated by CXR
brasfield score, mean FEV1, and FEF25%–75%, 21.55,
89.7%, and 75.9%, respectively. CXR was done only at
baseline to further evaluate disease severity and as part of
routineCFcare.PatientsrandomizedtoTSIwereyounger
and had milder lung disease than those randomized to
Placebo, although these differences were not statistically
significant. The difference in mean weight between the
two groups at baseline was statistically significant at
P¼0.04.
Mean changes in HRCT component scores between
treatment arms from visit 1 to visit 2 were analyzed for
consistency with the global total HRCT score and the
observed change in pulmonary function. Six of the eight
HRCT component scores were positively correlated with
the global total score (r>0.66); however, atelectasis and
inhomogeneity were not strongly correlated (r¼0.42 and
r¼0.50 respectively) and were observed to be highly
discordant with the observed changes in the global total
HRCT score, the HRCT component scores, and the PFT
changes.Inhomogeneitywasscoredonexpiratoryimages.
A modified total HRCT score was calculated by dropping
these two discordant component categories from the
overall total score (Table 2), resulting in a larger observed
HRCT treatment effect in the TSI group. The modified
total score decreased ?6.68?3.04 (mean?SD) for TSI
subjects and increased 0.02?2.0 (mean?SD) for the
placebo subjects (P¼0.07) (Table 3). We calculated
sample sizes that would be required to show statistical
significance using a modified total HRCT score, a global
total HRCT score, FEF25%–75%, or FEV1. A total of 60,
TABLE 1—Baseline Characteristics of Study Population
TOBI PlaceboP-value*
Number of patients
Gender (M/F)
Age (years)
Weight (kg)
Weight for age (percentile)
FEV1%predicted
FEF25%–75%predicted
HRCT global total score
CXR brasfield score
Sweat chloride (meg/L)
Genetic analysis
Homozygous delta F508
Heterozygous delta F508
16
6/10
16
6/10
11.81?7.46** (6.0–34.7)***
34.83?14.08 (17.6–62.0)
25%
95.73?17.21 (55.0–134.1)
81.71?22.94 (29.35–111.83)
27.77?17.64 (10.00–78.00)
22.05?1.32 (18.8–24)
101.1?19.84 (70.0–136.6)
15.86?7.25 (7.4–28.8)
46.38?16.92 (21.9–73.3)
10%
83.71?21.07 (45.00–108.73)
70.04?31.19 (17.36–110.14)
32.80?18.24 (6.25–76.25)
21.05?2.88 (14.8–25)
106.8?18.03 (80.0–145.0)
0.13
0.04
0.09
0.24
0.42
0.22
0.58
8
2
11
2
*Student’s t-test for two independent samples 95% level of significance (two-tailed).
**Mean?SD.
***Range.
1132Nasr et al.

Page 5

100, 200, and over 800 patients, respectively would be
required to demonstrate significance in this patient
population (Table 4).
Theinter-observercoefficientofvariationfortheglobal
total HRCT score was 19.9% (P¼0.000). Inter-observer
differences in reading scores is observed in about 1/3 of
the total (Table 6).
The intra-observer difference in the mean HRCT
reading for the first reader was P¼0.260 and for the
second reader, the P-value was 0.087. Non-parametric
tests(Wilcoxonexact)wereperformedforeachreaderand
for both readers combined (Table 5). The results of this
analysis were consistent with the t-test results. Table 5
showed significance only for reader 2 for atelectasis with
student’s t-test 0.027 and Wilcoxon 0.020. However, it
was noted that 97% of the observations had a maximum
scoredifferenceof1orless(maximumscore¼maximum
of the four reading scores?minimum of the four reading
scores). Interestingly, the interval between the two
readings for the first reader was 1–2 weeks while the
interval for the second reader was 1–2 months. A
frequency distribution of HRCT score differences was
also conducted for the 80 individual scores, by patients
and by case over the twovisits, to demonstrate the pattern
of scoring for each of the two readers. Ninety-seven
percent of the observations had a maximum score
difference of 1 or less (maximum score¼maximum of
the four reading scores?minimum of the four reading
scores). We analyzed the quality of life data (CFQ). No
useful results were obtained from the analysis due to the
small number of patients in each age group.
DISCUSSION
PFTsareconsideredthestandardforassessmentoflung
disease in CF. However, conventional PFTs are not very
sensitive in the detection of early lung damage.16In
addition, lung function tests are an indirect measure of
structural integrity of the airways.17HRCTwas shown to
be a more sensitive measure of detecting structural
abnormalitiesthanCXRinarecentstudy.4HRCTshowed
thatairwaysundergosubstantialthickeningearlyinlifein
CF patients.4HRCT has the potential to be a useful
outcome measure in CF. A necessary attribute of an
outcomemeasureisthatitimprovesrapidlywitheffective
therapy.7Thisisasecondstudythatwasconductedbyour
group to evaluate the value of HRCTas an end-point in a
double blind placebo controlled study.4In this study, we
demonstrated that HRCT could be a more sensitive
objective measure than PFTs in detecting a treatment
effect in CF pulmonary disease.
Previous studies evaluated the usefulness of HRCT in
understandingthenaturalhistoryofdifferentlungdisease.
In asthma, a study was conducted comparing airtrapping
as quantified by HRCT of the chest compared to lung
functionandairwayinflammationinchildrenwithmildto
moderate disease.18The data suggested that the quanti-
tative HRCT may be a useful tool in the evaluation of
peripheral airflow obstruction in children with asthma.18
A second study showed that patients with chronic
obstructive pulmonary disease (COPD) have more
prominentHRCTfindingsascomparedwithasthmatics.19
In another study, HRCT provided a complete and precise
assessment of children with bronchiectasis.20Ventilation/
perfusion scans and lung functions were found to be
additivetoolstotheunderstandingofthecomplexityofthe
disease process.20Several studies were done in CF to
evaluatetheseverityandprogressionoflungdiseaseusing
HRCT.2,7,13,21,22Other studies have evaluated early lung
disease in CF using HRCT.14,17,23,24These studies
addressed the natural history of the pulmonary disease
rather than an intervention. Our study investigated HRCT
asanoutcomesurrogateinashort-terminterventionstudy.
UsingHCRTtomeasuretreatmentoutcomesinpatients
with mild to moderate disease could result in smaller
number of patients required to demonstrate a significant
TABLE 2—Changes in HRCT Individual Category Scores Between Visits 1 and 2 and the Difference in Mean Change
Between TSI and Placebo Groups
HRCT Scores
TSI mean score
TSI mean
score change
Placebo mean score
Placebo mean
score change
Difference in mean
score change
Visit 1Visit 2 Visit 1Visit 2TSI?Placebo
4.23
0.59
1.70
0.04
?0.48
0.79
1.03
?0.77
2.44
Global total score
Bronchiectasis
Peribronchial thickening
Mucous plugging
Atelectasis/consolidation
Nodular opacities
Linear opacities
Inhomogeneity
Hyperinflation
27.78
4.03
7.95
0.30
0.34
2.04
2.93
3.63
6.57
23.72
3.48
6.71
0.39
0.48
1.11
1.98
6.11
3.45
4.06
0.55
1.24
(0.09)
(0.14)
0.93
0.95
(2.48)
3.12
32.80
5.41
9.08
0.56
0.47
2.20
3.25
7.24
4.74
32.63
5.45
9.54
0.69
0.13
2.06
3.33
8.95
4.06
0.17
(0.04)
(0.46)
(0.13)
0.34
0.14
(0.08)
(1.71)
0.68
HRCT in Evaluating TSI in Cystic Fibrosis 1133

Page 6

treatmenteffect.Ourcalculation showedthat100patients
wouldbesufficienttodemonstratestatisticalsignificance,
comparedto200andover800patientsifweusedFEF25%–75%
and FEV1 respectively. When the HCRT score was
modified by elimination of the two discordant scores
(inhomogeneity and atelectasis) from the global total
score, the number of patients needed to show statistical
significance decreased to 60 patients. Enrolling smaller
number of patients would allow faster completion of
studies, decreased numbers of patients exposed to overall
radiation, and decreased costs in conducting studies. In
addition, fewer patients would be subjected to investiga-
tional medications.
Peribronchial thickening, linear opacities, nodular
opacities, hyperinflation, and to a lesser degree, bronch-
iatasis and mucus plugging, were the most concordant
with the global total score (Table 2), and observed
improvement in PFTs. In a previous study, changes in
peribronhial thickening or airway thickening correlated
significantly with the change in FEF25%–75%predicted, a
lungfunctiontestbelievedtobesensitivetoabnormalities
in the peripheral airways.21Previous studies have shown
that bronchiatasis and mucus plugging were most
sensitive components of CF lung disease.21,22However,
these studies were evaluating progressive damage and
longitudinaldeclineinlungdisease.Ourstudywasdonein
mild–moderate CF patients; for this patient population, a
modified HRCT score might be more sensitive.
A trend towards improvement was noted in the HRCT
score comparedtobaseline intheTSI group,althoughnot
statistically significant, while no change was observed in
the placebo group (Table 2). Axial HRCT images in
TABLE 3—Mean Changes in FEV Predicted, FEF25%–75%, Total HRCT Global and Modified
Global Scores From Baseline to End of Study
TSI Placebo
FEV predicted
Visit 1
Visit 2
P-value
FEF25%–75%predicted
Visit 1
Visit 2
P-value
Global HRCT total score
Visit 1
Visit 2
P-value
Modified HRCT total score
Visit 1
Visit 2
P-value
95.73?17.21*
97.03?16.60
(55.0, 134.1)**
(66.17, 134.94)
83.71?21.07
84.46?20.85
(45.00, 108.73)
(40.23, 107.68)
0.97
81.71?22.94
87.78?24.77
(29.35, 111.83)
(31.17, 122.56)
0.04?31.19
68.16?30.66
(17.36, 110.14)
(18.30, 99.58)
0.23
27.78?17.64
23.72?12.02
(10.0, 78.0)
(9.0, 52.5)
32.80?18.23
32.63?17.65
(6.25, 76.25)
(7.25, 67.75)
0.13
23.81?14.54
17.13?8.86
(6.75, 60.5)
(8.5, 40.75)
25.24?17.22
25.26?14.91
(5.00, 73.25)
(6.75, 63.5)
0.07
*Mean?SD.
**Range.
TABLE 4—Sample Sizes Required to See Statistically Significant Differences Between Mean Changes in TSI and Placebo
Groups
Mean change
Visit 1?Visit 2
(?95% CI)
Sample sizes
31 60100 200800
Modified HRCT total score
TSI
Placebo
Global HRCT total score
TSI
Placebo
FEF25%–75%predicted
TSI
Placebo
FEV predicted
TSI
Placebo
?6.68(?13.27, ?0.09)
(?4.22, 4.26)
(?11.1,?2.26)
(?2.72, 2.76).02
?4.06
?.17
(?10.32, 2.21)
(?1.75, 5.24)
(?7.60, ?0.51)
(?0.17, 3.66)
6.08
?0.6
(3.43, 15.59)
(?5.19, 3.98)
(2.27, 0.88)
(?2.38, 0.17)
1.29
1.17
(?5.25, 7.83)
(?1.58, 3.92)
(?0.02, 2.60)
(0.63, 1.70)
1134 Nasr et al.

Page 7

inspiration at two levels from a patient in the TSI group at
baselineandendofstudyareshowninFigure1.Usingthe
modified HRCT total score in this mostly mild–moderate
CF lung disease population seems reasonable. This post
hoc analysis is potentially useful in developing an
improved measurement tool, although larger prospective
studies are needed to further evaluate this. The coefficient
of variation is considered by some to be a useful single
measure of the amount of variation in a distribution
relative to the magnitude of the mean of the distribution
and is often expressed as a percentage/ratio. We used the
formula described in Nasr et al.4The coefficient of
variation forthe HRCT global scorewas 19.9%. That was
lower than that found in Nasr et al. (49.1%),4but higher
than the 5% reported by Santamaria, et al.14The higher
range variation in HRCT scores in the Nasr study was
attributedbothtothesmallrangeofthescoringsystemand
the small number of patients (n¼12).4The higher inter-
observer coefficient of variation (19.9%) observed in this
study probably reflects the small number of patients and
the range of scoring scale for HRCT. One would hope to
find no differences in the mean scores either within each
reader or between the two readers. The coefficient of
variation is probably the most sensitive to any difference
between the raters on individual cases. A mean is a
summary measure, and can therefore mask differences at
theindividuallevel.Inaddition,tworeadersscorescanbe
highlycorrelatedandstillvarygreatly,forexample,ifone
rater’s scores are always exactly twice those of a second
rater, they will have a correlation of 1.0, with the best fit
line between their scores having a slope of two. Their
coefficientofvariationwillstillbehigh.Theintraobserver
variability showed lower reproducibility with longer
intervals (1–2 months) between the two readings. This
is consistent with other studies.5
Fig. 1. A–D: Girl (13-year-old) from the TSI group. Axial HRCT
imagesin inspiration obtainedbefore therapy (A,B) and 4 weeks after
starting therapy (C, D). On the initial examination (A, B), scattered
nodular opacities are seen within both lungs (arrows), more on the
right. Bronchial walls appear mildly thickened diffusely. On the follow-
up examination (C, D), the nodular opacities have resolved and the
bronchial walls do not appear as thick. Note slight inhomogeneity of
aeration, persisting after therapy (C).
TABLE 5—Summary of Changes in HRCT Score From V1–V2 by Each Reader and by Both Readers Combined
Components
Reader 1Reader 2 Both readers combined
Student’s t-test
Wilcoxon exact
P-valueStudent’s t-test
Wilcoxon exact
P-valueStudent’s t-test
Wilcoxon exact
P-value
Bronchiectasis
Peribronchial thickening
Mucous plugging
Atelectasis
Nodular opacities
Linear opacities
Inhomogeneity
Hyperinflation
Global
0.280
0.054
0.896
0.861
0.575
0.154
0.639
0.627
0.260
0.226
0.079
0.692
0.148
0.760
0.331
0.428
0.264
0.154
0.605
0.113
0.894
0.027
0.353
0.078
0.965
0.270
0.087
0.773
0.219
0.482
0.020
0.206
0.066
0.661
0.451
0.342
0.288
0.044
0.875
0.294
0.394
0.058
0.841
0.373
0.112
0.267
0.121
0.423
0.090
0.404
0.147
0.452
0.268
0.227
TABLE 6—Maximum–Minimum Difference Frequency
Check
The FREQ procedure
Difference FrequencyPercent
Cumulative
frequency
Cumulative
percent
0
1
2
3
4
3,138
1,727
135
62.65
34.48
2.70
0.12
0.06
3,138
4,865
5,000
5,006
5,009
62.65
97.13
99.82
99.94
100.00
6
3
Frequency missing¼31
HRCT in Evaluating TSI in Cystic Fibrosis 1135

Page 8

There are several limitations of this study. The first is
that it is a pilot. A larger study would be of interest to
further evaluate the usefulness of this measure in CF
patients with mild to moderate lung disease. Patients
randomized to TSI were younger and had milder lung
disease; however, there was no statistical significance
between the two groups.
AsecondlimitationistheuseoflimitedcutsoftheHRCT
whichreducesthesensitivitytodetectchanges.Thelimited
number of cuts could have also contributed to the high
interobserver coefficient of variation. However, using
limited cuts reduces radiation dosing that patients are
exposed to. In this protocol,therewere specific instructions
onhowandwheretoplacethefiveslicesatbaselineandend
of study. This should result on each slice being in the same
vicinity of the previous one, but not the exact position. We
felt that the five scans during inspiration and five during
expiration are reasonably representative of the disease
present and changes that occurred at the end of study. The
effective and equivalent doses of radiation from HRCT in
this study was approximately equal to two CXRs (PA and
lateral view each), even though the radiation dose is
delivered in a different manner for both procedures
(effective dose was 0.04 and 0.02 mSv and the equivalent
doses were 0.09 and 0.04 mGy respectively).
Athirdlimitationisthemethodofevaluatingexpiratory
images. We did not control for lung volume during
scanning as done in another study.23Although landmarks
were used to identify levels, it was difficult to exactly
match imaging levels from examination to examination.
Findings on expiratory images (hyperinflation and
inhomogeneity) were dependent on the patient following
breath holding instructions closely. A suboptimal expira-
tory effort could mimic hyperinflation or mask inhomo-
geneity. The importance of monitoring volume has been
reported by Robinson, et al.13Lack of volumemonitoring
may have increased the variability of this potentially
usefulcomponent.Furtherstudiesareneededwithvolume
monitoring to further analyze the value of measuring
inhomogeneity in this patient population.
Afourthlimitationwas ourinabilitytoreliablyanalyze
theCFQdata.Patientreportedoutcomesandqualityoflife
measures are increasingly being recognized as important
parameters in measuring treatment effect. Larger studies
are needed to evaluate CFQ value as an end point in mild
CF lung disease.
CONCLUSION
. HCRT can be a sensitive measure of change in mild to
moderate CF pulmonary disease.
. HRCT is likely to require a smaller sample size
than FEV1or FEF25%-75%to demonstrate a significant
treatment effect in patients with mild to moderate lung
disease.
. Atelectasis and inhomogeneity may not be useful
component scores in evaluating HRCT in mild to
moderate CF lung disease.
. The limited number of slices with low mA and/or scan
time reduces the radiation dose.
REFERENCES
1. Burns JL, Ramsey BW, Smith AL. Clinical manifestations and
treatment of pulmonary infections in cystic fibrosis. Adv Pediatr
Infect Dis 1993;8:53–66.
2. Dorlo ¨chter L, Nes H, Fluge G, Rosendahl K. High resolution CT
in cystic fibrosis—The contribution of expiratory scans. EJR
2003;47:193–198.
3. Mayo JR, Jackson SA, Mu ¨ller NL. High resolution CT of the
chest: radiation dose. AJR 1993;160:479–481.
4. Nasr SZ, Kuhns LR, Brown RW, Hurwitz ME, Sanders GM,
Strouse PJ. Use of computerized tomography and chest X-rays in
evaluating efficacy of aerosolized recombinant human DNase in
cystic fibrosis patients younger than age 5 years: A preliminary
study. Pediatr Pulmonol 2001;31:377–382.
5. de Jong PA, Ottink MD, Robben SG, Lequin MH, Hop WCJ,
Hendricks JJE, Pare ´ PD, Tiddens HAWM. Pulmonary disease
assessment in cystic fibrosis: Comparison of CT scoring systems
and value of bronchial and arterial dimension measurements.
Radiology 2004;231:434–439.
6. Konstan MW, Byard PJ, Hoppel CL, Davis PB. Effect of high-
dose ibuprofen in patients with cystic fibrosis. N Engl J Med
1995;332:848–854.
7. Brody AS, Molina PL, Klein JS, Rothman BS, Ramagopal M,
Swartz DR. High-resolution computed tomography of the chest in
children with cystic fibrosis: Support for use as an outcome
surrogate. Pediatr Radiol 1999;29:731–735.
8. Ramsey BW, Pepe MS, Quan JM, Otto KL, Montgomery AB,
Williams-Warren J, Vasiljev K-M, Borowitz D, Bowman CM,
Marshall BC, Marshall S, Smith AL. Intermittent administration
of inhaled tobramycin in patients with cystic fibrosis. N Engl J
Med 1999;340:23–30.
9. Murphy TD, Anbar RD, Lester LA, Nasr SZ, Nickerson B,
VanDevanter DR, Colin AA. Treatment with tobramycin solution
for inhalation reduces hospitalizations in young CF subjects with
mild lung disease. Pediatr Pulmonol 2004;38:314–320.
10. Quittner AL, Sweeny S., Watrous M, Munzenberger P, Bearss K,
Nitza AG, Fisher LA, Henry B. Translation and linguistic
validation of a disease-specific quality of life measure for cystic
fibrosis. J Pediatr Psychol 2000;25:403–414.
11. Brasfield D, Hicks G, Seng-Jaw S, Tiller RE. The roentgenogram
in cystic fibrosis: A new scoring system. Pediatrics 1979;63:24–
29.
12. Robinson TE, Leung AN, Northway WH, Blankenberg FG, Chan
FP, Bloch DA, Holmes TH, Moss RB. Composite spirometric-
computed tomography outcome measure in early cystic fibrosis
lung disease. Am J Respir Crit Care Med 2003;168:588–593.
13. Robinson TE, Leung AN, Northway WH, Blankenberg FG, Bloch
DA, Oehlert JW, Al-Dabbagh H, Hubli S, Moss RB. Spirometry-
triggered high-resolution computed tomography and pulmonary
function measurements during an acute exacerbation in patients
with cystic fibrosis. J Pediatr 2001;138:553–559.
14. Santamaria F, Grillo G, Guidi G, Rotondo A, Raia V, de Ritis G,
Sarnelli P, Caterino M, Greco L. Cystic fibrosis: When should
high-resolution computed tomography of the chest be obtained?
Pediatrics 1998;101:908–913.
15. Shah RM, Sexauer W, Ostrum BJ, Fiel SB, Friedman AC. High-
resolution CT in the acute exacerbation of cystic fibrosis:
1136Nasr et al.

Page 9

Evaluation of acute findings, reversibility of those findings, and
clinical correlation. AJR 1997;169:375–380.
16. Becklake MR, Permutt S. Evaluation of tests of lung function for
‘‘screening’’ for early detection of chronic obstructive lung
disease. In: The lung in the transition between health and disease.
New York, NY: Marcel Dekker, 1979.
17. Tiddens HAWM. Detecting early structural lung damage in cystic
fibrosis. Pediatr Pulmonol 2002;34:228–231.
18. Jain N, Covar RA, Gleason MC, Newell JD, Gelfand EW, Spahn
JD. Quantitative computed tomography detects peripheral airway
disease in asthmatic children. Pediatr Pulmonol 2005;40:211–
218.
19. Harmanci E, Kebapci M, Metintas M, Ozkan R. High-resolution
computed tomography findings are correlated with disease
severity in asthma. Respiration 2002;69:420–426.
20. Pifferi M, Caramella D, Bulleri A, Baldi S, Peroni D, Pietrobelli
A, Boner AL. Pediatric bronchiectasis: Correlation of HRCT,
ventilation and perfusion scintigraphy, and pulmonary function
testing. Pediatr Pulmonol 2004;38:298–303.
21. de Jong PA, Nakano Y, Lequin MH, Mayo JR, Woods R, Pare ´ PD,
Tiddens HA. Progressive damage on high resolution computed
tomography despite stable lung function in cystic fibrosis. Eur
Respir J 2004;23:93–97.
22. Brody AS, Molina PL, Klein JS, Campbell JD, Millard SP, Quan
J. High-resolution CTis more sensitive to longitundinal decline in
lung status in young children with CF than pulmonary function
tests. Pediatr Pulmonol 2003;36:A388.
23. Long FR, Williams RS, Castile RG. Structural airway abnorm-
alities in infants and young children with cystic fibrosis. J Pediatr
2004;144:154–161.
24. Bonnel A-S, Song SM-H, Kesavarju K, Newaskar M, Paxton CJ,
Bloch DA, Moss RB, Robinson TE. Quantitative air-trapping
analysis in children with mild cystic fibrosis lung disease. Pediatr
Pulmonol 2004;38:396–405.
HRCT in Evaluating TSI in Cystic Fibrosis1137