Young children have greater susceptibility to airway narrowing leading to air trapping or atelectasis than older children and adults. This is attributed to young children having higher airway resistance, weaker airway skeletons, and more abundant bronchial mucous glands than older patients. When mosaic lung attenuation is seen on inspiratory CT, expiratory CT can be used to differentiate air trapping from ground-glass opacity or vascular obstruction [1, 2]. However, expiratory CT is difficult to perform on young children who cannot hold their breath. Nevertheless, there have been attempts to obtain expiratory CT images from young children using cine CT [3-7], lateral decubitus CT [8-10], and controlled-ventilation CT [11, 12]. Cine CT studies have been performed using electron beam CT [3-6] and helical CT with 1.0-second gantry rotation time [7], but no studies have yet been conducted using helical CT with a faster gantry rotation time, to our knowledge. In this study, we evaluated whether 0.3-second cine CT can be used for assessing air trapping in the lungs of free-breathing young children.

Thirty young children (15 boys and 15 girls; mean age, 25 months; age range, 1 month-6 years) who were suspected of having lung areas with air trapping and who underwent cine CT studies were included in this study. This study was approved by the institutional review board, and informed parental consent was received before each CT examination.

Air trapping was considered to be present when the following imaging findings were seen on cine CT images. First, a hyperlucent area in the lung should be detected on CT images. Second, the pulmonary vessels in the hyperlucent area should be smaller than those in an area of normal lung. Third, the hyperlucent area should become less bright or even darker than an area of normal lung on expiratory phase CT images. Fourth, a change in the hyperlucent area size should be smaller than that in a normal lung area of the same size on cine CT images or the hyperlucent area should become larger on expiratory phase CT images than on inspiratory phase CT images.

Two radiologists visually assessed cine CT images by consensus and identified the presence of air trapping in 24 of the 30 children. Their diagnoses are shown in Table 1. On the other hand, six patients belonged to a no-air trapping group and their diagnoses are described in Table 2. Diffuse lung abnormalities were present in six patients. Among them, two patients showed diffuse hyperlucency and four patients showed diffuse opacity. We reviewed available 3D helical CT images obtained on the same day as cine CT or on another day close to the date of the cine CT examination to look for airway stenoses responsible for air trapping detected on cine CT.

A CT system (LightSpeed QX/i, GE Healthcare) with 0.5-second gantry rotation time and 4-slice capability was used to obtain 0.3-second cine CT studies. Each 1.25-mm slice was scanned for 3 seconds with 100 kVp and 50-90 mA at a fixed level while the patient was breathing quietly in the supine position. Cine CT was usually performed without sedating the patient because an IV contrast agent was not used and the scanning time of the study was short. Serial segment images were reconstructed retrospectively from CT data obtained with a 225° rotation (0.3 seconds) and an intersegmentation interval of 0.3 seconds. Sixty cine CT images per patient (six slice levels, 10 images at each level) were reconstructed using a standard algorithm. The high-spatial-frequency bone algorithm was not available for this segment reconstruction. Two radiologists in consensus selected inspiratory and expiratory phase images on the basis of lung volume, lung density, and chest wall motion by paging through the cine CT images.

Lung density was measured by placing a rectangular region of interest (6-29 mm² in area) over an area with abnormal density and an adjacent normal area (if present) on a pair of inspiratory and expiratory phase CT images, with careful attention not to include large pulmonary vessels (Figs. 1A and 1B). For the measurement of normal lung density, an area was chosen that did not show compressive atelectasis in the adjacent ipsilateral lung. Normal lung density could not be measured in the six patients with diffuse abnormalities. These patients were included in the analysis of abnormal lung density but were excluded in the analysis in which normal lung density was necessary.

Lung density differences between inspiration and expiration were calculated for all patients and compared in areas with air trapping, in abnormal areas without air trapping, and in normal areas. The density difference percentage ([density difference_abnormal/density difference_normal] × 100) between inspiration and expiration was calculated and compared between the group with air trapping (n = 22) and the group with no air trapping (n = 2) to compensate for individual tidal volume variation during quiet breathing that could potentially affect results of lung density measurements.

In addition, lung density differences between abnormal areas and adjacent normal areas were calculated at the inspiratory and expiratory phases and were compared between the two groups. Lastly, we evaluated whether mosaic lung attenuation due to air trapping was visible on inspiratory phase images in the 22 patients of the air trapping group. For all statistical comparisons, unpaired and paired Student's t tests were used and a p value of less than 0.05 was considered to indicate a statistically significant difference.

In all patients, we could successfully select inspiratory and expiratory phase images from cine CT images and successfully measured lung densities on the CT images. In 11 of the 24 patients with air trapping, we found central airway stenoses responsible for air trapping on 3D helical CT images, which was supplementary evidence for the presence of air trapping visually detected on cine CT. We could not delineate peripheral airway stenoses because of the limited spatial resolution of the CT images.

For all patients, lung density differences between the inspiratory and expiratory phases were significantly smaller in areas with air trapping (mean ± SD, -19 ± 34 H) than in abnormal areas without air trapping (138 ± 36 H) (p < 0.001) or normal areas (111 ± 49 H) (p < 0.001) (Fig. 2). Among six patients with diffuse abnormalities, two patients with diffuse air trapping (both from the air trapping group) could be objectively distinguished from four patients with diffuse increased opacity (all from the no-air trapping group) on the basis of CT densitometry (Figs. 3A, 3B, and 3C). Density difference percentages were significantly smaller in the air trapping group (-27% ± 54%) than in the no-air trapping group (120% ± 87%) (p < 0.001) (Fig. 4).

In the air trapping group, lung density differences between abnormal and normal areas were significantly larger at the expiratory phase (260 ± 77 H) than the inspiratory phase (129 ± 69 H) (p < 0.001), but did not change significantly through the respiratory cycle in the no-air trapping group (213 ± 186 H and 207 ± 105 H, respectively) (p > 0.05) (Fig. 5). Of the 22 patients with focal air trapping detected on expiratory phase images, mosaic lung attenuation due to air trapping was barely visible in six patients (27%) on inspiratory phase images (Fig. 6A and 6B).

Expiratory CT has been reported to be useful in identifying mosaic lung attenuation due to air trapping, which indicates the presence of large or small airways disease [1, 2]. Air trapping is more prevalent in young children than older children and adults, and expiratory CT is therefore of significant clinical importance for this group. However, breath-hold expiratory CT examinations cannot be performed on young children who cannot hold their breath. To overcome this difficulty, cine CT using electron beam CT [3-6] or helical CT [7], lateral decubitus CT [8-10], and controlled-ventilation CT [11, 12] have been performed to obtain expiratory CT images in young children. To our knowledge, cine CT with a 0.3-second temporal resolution has not yet been evaluated for detecting air trapping in young children.

In this study, air trapping was accurately detected in 24 free-breathing young children using cine CT. In addition to focal air trapping, diffuse air trapping was detectable in two patients with objective measurements of lung density on cine CT, thereby confirming the findings of other authors that expiratory CT can be helpful in the diagnosis of diffuse lung disease [2]. In addition to lung density assessments, changes in lung volume, the chest wall, and cardiomediastinal structures through the respiratory cycle may be other helpful clues in identifying focal or diffuse air trapping. These changes tend to be restricted in cases of large areas of air trapping and are prone to being exaggerated in adjacent areas of normal lung.

In six (27%) of 22 patients with mosaic lung attenuation, we detected air trapping on only the expiratory phase CT images, exemplifying the additional usefulness of expiratory CT. Other researchers have also detected air trapping using expiratory CT in patients who had normal findings on inspiratory CT [13]. Those authors found that the patients showed pulmonary function test results that were intermediate values between the results of patients with normal findings on inspiratory and expiratory CT scans and those of patients with air trapping and abnormal findings on inspiratory CT, suggesting a mild degree of air trapping [13].

Cine CT can be performed using electron beam or helical CT. Although temporal resolution of electron beam CT is higher than that of helical CT, helical CT with subsecond gantry rotation time, which was used in our study, is more readily available. In cine CT, temporal resolution is reduced further to approximately half of the gantry rotation time by the use of partial reconstruction. Higher temporal resolution may be beneficial in obtaining more accurate lung density measurements by reducing respiratory and cardiac motion artifacts. Based on our previous experiences [9], lateral decubitus CT may be somewhat limited owing to the necessity of patient position changes during the procedure. We found that this occasionally caused sedation failure and remarkable respiratory misregistration between the two lateral decubitus CT images at a defined slice position, making lung densitometry difficult. Another alternative way to perform cine CT that is safe and provides excellent image quality is controlled-ventilation CT [11, 12], in which the CT scan is obtained during respiratory pauses induced by synchronous controlled positive airway pressure delivered with the Sellick maneuver. However, this method is complex and requires two experienced operators and a special device for controlled ventilation. Cine CT, as used in our study, avoids these problems because it can be performed with the patient in the supine position and requires no special devices, maneuvers, or patient sedation.

Another important issue in imaging pediatric patients with CT is radiation dose. The CT dose index volume (CTDI_volume) of cine CT was 6.6-11.8 mGy at one level in our study. Therefore, the calculated dose-length product at all six levels was approximately 5.0-8.9 mGy × cm, and the estimated effective dose [14] was 0.09-0.15 mSv. Hence, the radiation dose of our cine CT protocol was relatively low compared with that of other chest CT methods. We believe that acquiring cine CT images at six levels may be sufficient for identifying significant air trapping. Detection of a lobular or subsegmental area of air trapping may not be necessary because it can be seen in healthy subjects and may be clinically insignificant [15]. To further reduce radiation dose, we propose reducing the number of levels to encompass abnormal lung areas only. Despite the relatively low radiation dose of cine CT, we believe that cine CT should be performed for the appropriate indications to maximize clinical benefits from the study.

One of several limitations of this study is a lack of standard reference for the diagnosis of air trapping. However, the pulmonary function test is difficult to perform in young children, and the test results may be misleading in uncooperative patients. Moreover, detection of air trapping based on CT densitometry has already been assessed in many cine CT studies [3-7]. In the expiratory phase, the proportion of air having very low CT density in a lung volume is relatively high in areas of air trapping compared with normal lung areas. The principle of performing CT densitometry for detecting air trapping has been used in previous cine CT studies [3-7], and it also was proved to be accurate and objective in this study. Regarding comparison between the two groups in this study, the small number of children in the no-air trapping group is another limitation of our study. However, it would not be ethical to perform cine CT in children who were not suspected of having air trapping in their lungs. Like other pathologic conditions, air trapping may be a spectrum of abnormality. Therefore, subtle or mild air trapping may be missed and cannot be differentiated from a normal lung area by means of qualitative and quantitative evaluations of lung density changes on cine CT. This limitation may cut down the sensitivity of cine CT in detecting air trapping, but mild air trapping missed on cine CT may be less significant clinically than moderate or severe air trapping.

Density differences of a normal lung between inspiration and expiration have been reported to be in the range of 100-150 H [3, 16]. Density differences in normal lung areas (mean, 111 H) and in abnormal areas without air trapping (138 H) of this study fell within that range. In a previous study using lateral decubitus CT [9], a mean density difference between normal lung areas and areas of air trapping in dependent lungs (208 H), equivalent to expiratory phase imaging, and that in nondependent lungs (121 H), equivalent to inspiratory phase imaging, were also similar to results from this study (260 and 129 H, respectively).

Of the 24 patients with air trapping in our study, 46% of the cases (11/24) were caused by airway stenoses that were apparent on 3D CT images of the airways. Air trapping in the remaining 13 patients might have been due to small airways disease, and corresponding airway abnormalities might not, therefore, be directly identified on CT images. We think that the calculation of density difference percentages is not necessary for detecting air trapping on cine CT because the results of density difference percentages showed the same statistical significance as those of density differences.

In conclusion, air trapping can be accurately detected in the lungs of free-breathing young children using 0.3-second cine CT regardless of whether it is focal or diffuse in extent. Detection is possible because an area with air trapping shows a smaller increase or a paradoxical decrease in lung density at the expiratory phase.

Address correspondence to H. W. Goo ([email protected]).