| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, Section of Paediatrics, Haukeland University
Hospital, 5021 Bergen, Norway.
2 Department of Paediatrics, Haukeland University Hospital, Bergen,
Norway.
3 Department of Public Health and Primary Health Care, and Medical Birth
Registry of Norway, Section for Epidemiology and Medical Statistics,
University of Bergen, Bergen, Norway.
Received March 4, 2005;
accepted after revision September 4, 2005.
Address correspondence to S. M. Aukland
(stein.magnus.aukland{at}helse-bergen.no).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. High-resolution CT of the lungs was assessed
in 72 children born at a gestational age of 
28 weeks or with a birth
weight of 1,000 g within a defined region in western Norway in 1982-1985
(n = 40) or in 1991-1992 (n = 32). All images were analyzed
by two pediatric radiologists using a scoring system in which a total of 14
features were assessed.
RESULTS. Sixty-three (88%) of the subjects had abnormal findings,
the most common being linear opacities (n = 52), triangular opacities
(n = 42), air trapping (n = 19), and mosaic perfusion
(n = 10). Right and left lungs were equally affected. There were
fewer abnormalities in the younger age group (born in 1991-1992).
Intraobserver agreement and interobserver agreement were moderate (weighted
= 0.54 and = 0.52, respectively). Fifty-six of the 72 children
had a clinical diagnosis of bronchopulmonary dysplasia, and the median total
score and the median scores of the four most common findings were higher in
the bronchopulmonary dysplasia group; however, the differences were not
statistically significant.
CONCLUSION. High-resolution CT in young people of preterm birth revealed abnormal radiologic findings in 81.3% of the patients at age 10 years and 92.5% at age 18 years. Linear, triangular, and subpleural opacities were the most common. The reproducibility of the applied scoring system was acceptable.
Keywords: bronchopulmonary dyplasia • high-resolution CT • lung disease • neonatal imaging • pediatric radiology
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The radiographic pulmonary findings in extremely preterm infants have changed [5]. Traditional radiologic findings in BPD survivors in middle childhood included fibrosis, patchy atelectasis, and emphysema. Newer findings are discrete perihilar lung opacification in the neonatal period that apparently normalizes during childhood. However, the sensitivity of radiography in the diagnosis of subtle lung abnormalities is low, and the number of indications for high-resolution CT (HRCT) of children has increased [6]. Oppenheim and coworkers [7] performed pulmonary HRCT on 23 selected BPD survivors born in the late 1980s and described radiologic findings for all patients. To our knowledge, results from population-based, unselected cohorts of preterm infants have not been presented, and our knowledge on the long-term effects of changes in perinatal and neonatal care is scarce. Systems for classification of findings on pulmonary HRCT in a population of this kind are not readily available.
The purpose of this study was to examine structural lung sequelae after extremely preterm birth by applying a novel HRCT scoring system to two population-based cohorts, one born in the early 1980s and one born in the early 1990s. We also assessed intra- and interobserver variability of the scoring system.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
28 weeks or with a birth weight of 1,000 g during
the periods 1982-1985 (first cohort) and 1991-1992 (second cohort) within a
defined region in western Norway were examined. Medical care was provided at
the only regional neonatal intensive care unit in the region, Haukeland
University Hospital, by senior medical staff who were essentially similar in
the two periods. Preterm infants were considered enrolled when admitted to the
neonatal department. Eighty-six (66%) of 130 preterm infants were alive at the
follow-up examination. Five eligible preterm infants were inaccessible. Hence,
81 (94%) of the survivors agreed to enter the study, 46 born in the 1980s and
35 in the 1990s. As part of an extended follow-up program, HRCT of the lungs
was successfully performed on 74 (91%) of the subjects between March 2001 and
March 2002. One child was medically unable to undergo CT, and six children did
not arrive for the appointment. Two patients were examined, but the images
were printed on hard copy and accidentally not stored in the PACS. Thus images
of 40 young adults (mean age, 18 years) and 32 children (mean age, 10 years)
were assessed. The mean gestational age at birth was 27.2 weeks (SD, 1.5
weeks), and the mean birth weight was 991 g (SD, 191 g). Fifty-six (77.8%) of the 72 children were given the clinical diagnosis of BPD in the neonatal period. Neonatal data on the nine subjects who were not examined did not differ significantly compared with that on the patients examined. No subjects were examined within 2 weeks of a respiratory tract infection or an asthma event. The examinations were done without administering sedation. The regional ethics committee approved the study, and informed written consent was obtained from all subjects and their caretakers.
HRCT Imaging and Analysis
HRCT of the lungs was performed with a single-detector CT scanner (HiSpeed
Advantage, GE Healthcare) with 1.25-mm slice thickness, 0.5-second scanning
time, 120 kV, 50-100 mA, lung algorithm, and 512 x 512 matrix.
Approximately 10-12 scans in inspiration at 10-mm intervals were followed by
four or five scans in expiration at 20-mm intervals. Total radiation exposure
per examination was not measured directly, but dose equivalent was estimated
at 0.5-1.0 mSv per examination.
All images were displayed at a window width of 1,540 H and window level of -400 H at a PACS workstation. On the basis of the Bhalla score [8], we developed a scoring system that included 14 parameters (Table 1). On the basis of results of previous studies of children with BPD [7, 9], we added the findings of mosaic perfusion on inspiration and air trapping on expiration. In addition, we scored the occurrence of linear opacities and triangular subpleural opacities and evaluated the width of the interlobular septa and whether the bronchial artery had a greater diameter than the accompanying bronchus (bronchus and accompanying pulmonary artery within 1 mm of each other). Evaluation of the extent of disease was based on a geographic lung map with lung segment borders in which all pathologic findings were recorded.
|
All parameters except number of bullae had a defined maximum score. The theoretic maximum score for these 13 elements was 152. The occurrence of bullae was low, and therefore this maximum score also was the practical maximum score. To our knowledge, no objective method of evaluating mosaic perfusion disturbances has been described. In the present study, we defined mosaic perfusion as hypoattenuated areas in the lung on inspiratory scans [10], including both hypoattenuated areas with small vessels (mosaic oligemia) and hypoattenuated areas with normal-caliber vessels. Evaluation of mosaic perfusion also was influenced by the window and level settings and was disturbed by motion artifacts. In cases showing inspiratory mosaic perfusion in one lung segment and hypoattenuation in the same area on the expiratory scan, we scored the findings separately as mosaic perfusion on inspiration and air trapping on expiration. Air trapping was defined as radiolucent or hypoattenuated areas on expiratory scans. The term "emphysema" was defined as an area of destroyed lung architecture.
The examinations were interpreted independently by two experienced pediatric radiologists. Observer 1 reviewed the HRCT images during a 3-month period in the spring of 2003 and reinterpreted the images after 6 months. Observer 2 interpreted the images once in May 2004. Neither of the observers knew previous results or clinical findings. Before the study, to standardize the scoring system, both observers analyzed and discussed the findings of four HRCT examinations (three patients with known cystic fibrosis and one who had been a premature infant, later included in the study).
Statistical Analysis
Differences in the occurrence of pathologic findings between girls and boys
and between the two age groups were tested with a nonparametric test
(Mann-Whitney). Agreement within and between observers (total HRCT score,
subtotal scores for linear opacities, triangular opacities, mosaic perfusion,
and air trapping) was examined with Bland-Altman plots and the kappa
statistic. Kappa values were interpreted according to Altman
[11] ( < 0.20, poor;
= 0.21-0.40, fair; = 0.41-0.60, moderate; = 0.61-0.80,
good; = 0.81-1.00, very good). Differences due to grading were tested
with the McNemar test of symmetry. All reported p values are
two-tailed, and p < 0.05 was considered statistically significant.
The Mann-Whitney U test was used to evaluate differences in score
between patients with and without BPD.
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
The mean total HRCT score was 6.9 (95% CI, 5.3-8.6). The mean scores for the four most common findings varied between 0.7 and 2.3 (Table 3). There were no differences in mean total HRCT score between girls and boys (5.5 vs 8.8; p = 0.10). Moreover, no differences were found between girls and boys in mean score for linear opacities, triangular opacities, mosaic perfusion, or air trapping. Children 10-11 years old had a lower total HRCT score than subjects 17-19 years old (5.7 vs 7.9, respectively; p = 0.019). Similar findings were seen for triangular opacities (mean HRCT score, 1.1 vs 2.3; p = 0.018) and for air trapping (0.8 vs 1.8; p = 0.027).
|
The 56 children who were judged to have BPD in the neonatal period had a higher mean and a higher median total score and a higher score for linear opacities, triangular opacities, mosaic perfusion, and air trapping; however, the differences were not statistically significant (Table 4).
|
Reproducibility of the Method
Intra- and interobserver agreement for total score and for the most common
radiologic findings is shown in Figures
4A,
4B,
4C,
4D,
4E,
5A,
5B,
5C,
5D, and
5E. For the total score, the
degree of agreement was higher for low-score cases than for high-score
cases.
|
|
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
A semiquantitative scoring system for assessment of pulmonary disease has proved useful, especially in the follow-up of patients with cystic fibrosis [12]. Various systems have been designed [8, 13-15] and when compared have turned out to be equally reliable and robust [7]. At present, HRCT is an important additional diagnostic tool in the follow-up of patients with cystic fibrosis and has great potential to become an outcome surrogate for lung disease in cystic fibrosis [16]. Having used the Bhalla scoring system for several years, we modified it for assessment of pulmonary findings in children who had been born prematurely by adding an evaluation of air trapping and subtle lung opacities. Kubota et al. [17] reported on an ultrafast CT scoring system for assessing BPD. This system, however, was designed for children with known BPD and did not include all pathologic findings of interest in our group of children.
The scoring system used in this study did not weigh pathologic findings according to degree of severity; thus the findings linear opacities and emphysema had a similar effect on the total HRCT score and as such represent a weakness of our scoring system. However, because the more common findings were linear and triangular opacities, air trapping, and mosaic perfusion and there were only a few cases of collapse or consolidation and no cases of emphysema, the total HRCT scores are likely to reflect the degree to which lung was affected in the present population. The clinical validity of our findings remains unclear, however, and should be addressed in future studies. Increasing knowledge of clinical validity should also make it possible to refine the scoring system by weighing the different HRCT findings. We did not correlate the HRCT findings to results of pulmonary function tests, and that is a weakness of our study.
Although children with a clinical diagnosis of BPD had overall higher CT scores than those without that diagnosis, the differences were not statistically significant. Several explanations have to be considered. First, the number of children without BPD in our cohort was low, decreasing the statistical power of the tests used. Second, the clinical criteria for BPD may be nonspecific, resulting in overlap between the groups. Finally, the radiologic finding noted may not have been associated with BPD in the neonatal period; this hypothesis is not supported by other authors [7, 9].
Linear and Triangular Opacities
Differentiation between linear opacities and small peripheral vessels was
difficult in interpretation of the images. Although the examinations were
performed with 1.25-mm thin sections, partial-volume effect may partly explain
the problem. Moreover, the interlobar septa have a triangular base at the lung
surface, which can be misdiagnosed as pathologic triangular opacity
[7]. In addition, the linear
and triangular opacities were often linked or in proximity
[7]. In these cases, we scored
them separately, with a calculated risk of overestimating the number of
findings.
The findings of linear opacities and triangular opacities as the major lung abnormality in these children are in accordance with findings in earlier publications [7, 9, 17]. The results of our study support the view that these findings are part of the pulmonary damage in patients surviving BPD. The prevalence rates of linear opacities and triangular opacities in our study were 80.6% and 58.5%, respectively; however, the mean scores were low. For children with a history of BPD, the corresponding figures were 80.4% and 58.9%. In comparison, Oppenheim and coworkers [7] found prevalence rates of 95.6% and 100%, respectively, in their study of 23 children with a history of BPD. We cannot explain the marked difference in prevalence rates in these two studies.
Mosaic Perfusion
The prevalence of hypoattenuation with mosaic perfusion in our study was
low, only 13.8%, compared with the prevalence of 80-90% reported by others.
Aquino et al. [18] reported
the CT abnormalities in 24 of 26 patients with a clinical history of BPD; 20
of these patients had areas of decreased lung attenuation, and 24 patients had
air trapping. Howling et al.
[9] found extensive bilateral
areas of reduced lung attenuation in all their patients when reviewing five
adult survivors of BPD and comparing the findings with those for 10 control
subjects. In the study by Oppenheim et al.
[7], 20 (87%) of 23 patients
had areas of abnormally low attenuation.
In our study, the finding of mosaic perfusion was consistent between and within observers, supporting the hypothesis that mosaic perfusion may be considered a sequela of BPD. Although the differences between the BPD group and the non-BPD group were not statistically significant, a near-significant p value (0.07) and mean scores of 0 and 0.95, respectively, strengthen our belief that mosaic perfusion may represent permanent lung disease.
Air Trapping
The finding of air trapping had a high degree of reproducibility and was
seen in 19 of the 72 patients. Eleven of the 19 patients had hypoattenuated
areas on expiratory scans only and normal attenuation on inspiratory scans.
Although the additional expiratory scans increased the radiation dose by a
factor of 1.5, the total dose was low. We believe the dose was justified by
the increased diagnostic yield. Other investigators
[18] have found higher rates
of air trapping, up to 92%; however, their population differed from ours.
Neither Oppenheim et al. [7]
nor Howling et al. [9] included
expiratory scans in their studies. All children with air trapping had
additional abnormal findings, and the mean total score for this group was
11.8, of which air trapping alone constituted 41% of the total score. The
clinical value of adding an expiratory scan remains unclear. In their study of
children with difficult-to-control asthma, Marchac et al.
[19] found no association
between air trapping on expiratory scans and clinical severity. Other authors,
however, report that expiratory scans are of value in the evaluation of cystic
fibrosis [20,
21].
Bronchiectasis
Evaluation of the severity and extent of bronchiectasis is flawed by
methodologic difficulties
[19], as illustrated by our
low agreement between and within observers for this finding. We found mild
bronchiectasis in 9.7% of our population, who had a median age of 15.7
years.
Bronchus-to-Bronchial Artery Diameter Ratio
Unlike Howling et al. [9],
we found no cases of decreased ratio between neighboring bronchus and
bronchial artery. We examined a large group and used an ambitious scoring
system, deciding not to use any calibration tools or systematic magnification
in the CT interpretation. These factors may explain part of the
difference.
Other Findings
The low prevalence of bullae, consolidation, emphysema, and mucus plugging
corresponds to results of earlier studies. There seems to be no association
between lung damage as part of BPD and these abnormal findings on HRCT.
Peribronchial thickening was found in four patients and was always combined
with bronchiectasis. Evaluation of thickening of the interlobar septa has, to
our knowledge, not been previously performed. Our high prevalence of
approximately 10% suggests that the thickening may be a lung change suggestive
of BPD.
Intraobserver Variability
We found moderate reproducibility for all CT scores, including linear
opacities, triangular opacities, and mosaic perfusion ( = 0.45). In
comparison, de Jong and co-workers
[12] found kappa values of
0.67, 0.45, 0.73, respectively. Recoding the score of the findings into one of
four categories (score of 0-5 = 1, 6-10 = 2, 11-15 = 3, 16-20 = 4) did not
improve reproducibility rate or interobserver agreement. For the four more
common findings, we also performed analyses with cutoff values of 3, 4, and 5.
That is, an HRCT score lower than the cutoff value was assigned a zero value,
and scores equal to or above the cutoff value were given a value of 1. These
interim analyses did not significantly improve the reproducibility rate.
Interobserver Variability
We found good agreement between observers for total HRCT score; however,
the degree of agreement varied substantially among parameters. After
categorizing the findings as normal or abnormal (present or not present), we
found concordance in 54 (75.0% of the total 72) cases of linear opacities, 57
(79.2%) cases of triangular opacities, 67 (93.1%) cases of mosaic perfusion,
and 63 (87.5%) cases of air trapping. These results are comparable to the
results of Kubota et al.
[17].
Conclusion
The population of infants who survive extreme immaturity is growing, and
follow-up studies including not only BPD survivors but also full preterm
cohorts are becoming more important. Lung imaging with HRCT in childhood and
early adulthood will supply baseline data in the long-term follow-up of this
population. We found no statistically significant correlation between the
radiologic findings and the clinical diagnosis of BPD in the neonatal period.
Assessment with a CT scoring system with moderate reproducibility showed a
large number of subjects in this study had abnormal lung HRCT scans. Most of
the abnormalities were discrete linear or triangular opacities together with
mosaic perfusion and air trapping. The clinical validity of these findings is
unclear and has to be addressed in future studies.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  S M Aukland, K Rosendahl, C M Owens, K R Fosse, G E Eide, and T Halvorsen Neonatal bronchopulmonary dysplasia predicts abnormal pulmonary HRCT scans in long-term survivors of extreme preterm birth Thorax, May 1, 2009; 64(5): 405 - 410. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. M. Wong, A. N. Lees, J. Louw, F. Y. Lee, N. French, K. Gain, C. P. Murray, A. Wilson, and D. C. Chambers Emphysema in young adult survivors of moderate-to-severe bronchopulmonary dysplasia Eur. Respir. J., August 1, 2008; 32(2): 321 - 328. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Baraldi and M. Filippone Chronic Lung Disease after Premature Birth N. Engl. J. Med., November 8, 2007; 357(19): 1946 - 1955. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
