Normal and abnormal dynamic appearances of the pediatric upper airway can be evaluated with 4D CT, which is one of the most common clinical applications in newborns and infants (91011121314). Generally, the uvula, tongue, epiglottis, vocal cord, and pyriform sinuses serve as anatomic landmarks for image evaluation. Upper airway dimension is variably distended during inspiration and variably collapsed during expiration, depending on the amount of airflow and the strength of the supporting soft tissue structures. Intermittent velopharyngeal closure separating the oral cavity from the nasal cavity may be observed on 4D CT as a normal dynamic finding (Fig. 2). Therefore, 4D CT has the potential to assess velopharyngeal insufficiency caused by various disorders such as cleft palate and neuromuscular disorders. In upper airway obstruction, 4D CT is particularly useful to simply distinguish a dynamic obstructive lesion (Fig. 3) from a fixed one (Fig. 4), which is crucial for treatment planning. Glossoptosis, defined as abnormal posterior displacement of the tongue and commonly seen in children with micrognathia (particularly in Pierre Robin syndrome), Down syndrome, and neurologic impairment, may cause oropharyngeal obstruction and obstructive sleep apnea. The diagnosis of glossoptosis (Fig. 5) can be made with neck lateral view, fluoroscopy, 4D CT, and endoscopy. Radiologists are usually not familiar with normal (Fig. 6) and abnormal laryngeal motion, and hence, a learning curve is necessary for 4D CT evaluation in patients with suspected laryngomalacia. Laryngomalacia, the most common cause of inspiratory stridor in the first year of life, is caused by excessive inward collapse of the supraglottic structures. The diagnosis and severity of laryngomalacia can be assessed with fiberoptic nasolaryngoscopy and 4D CT (Fig. 7). It is usually self-limited, but surgical treatment is necessary in a small subset of patients with severe laryngomalacia. 4D CT may also be preferably used to evaluate treatment response in laryngomalacia because of its noninvasive nature (Fig. 7).

Tracheobronchomalacia is defined as an excessive expiratory collapse of the tracheobronchial tree. In addition to the primary type, the secondary type of tracheobronchomalacia may develop as a result of several pathophysiologic mechanisms such as tissue injury, inflammation, and longstanding extrinsic compression. Tracheomalacia is commonly associated with tracheoesophageal fistula and therefore should be strongly suspected and evaluated for, in patients suffering from postsurgical respiratory symptoms. Paired end-inspiratory/end-expiratory CT (1), dynamic expiratory CT, cine CT (20), or the more recent 4D CT (91011121314) (Fig. 8) may be used for the diagnosis of tracheomalacia. Compared to other imaging methods, the evaluation with 4D CT can be performed during free breathing, regarded as a more physiologic state, without additional respiratory maneuvers or procedures. Moreover, its gapless 4D imaging capability can overcome the limitations of cine CT, including difficulties in evaluating longitudinally moving and obliquely coursing structures. As a result, 4D CT allows the accurate evaluation of the extent and severity of tracheomalacia, and it makes the diagnosis of bronchomalacia feasible. Traditionally, an empirical cut-off value of 50% in a reduction of cross-sectional area has been used for the diagnosis of tracheomalacia, not only in adults, but also in children (1). However, its diagnostic value is compromised by the fact that the tracheal collapsibility in normal subjects may also be greater than 50%. In a comparative study between 27 children with bronchoscopy-proven tracheomalacia and 320 control patients (20), a diagnostic cut-off value of 31.6% in tracheal collapsibility on free-breathing cine CT was recently suggested with 96.3% sensitivity, 97.2% specificity, and 97.1% accuracy for the diagnosis of tracheomalacia. In line with the result, tracheal collapsibility on free-breathing 4D CT was less than 34% in normal pediatric patients and in the range of 43–90% in patients with malacic airways (13). One of the current problems is that time-consuming manual airway tracing on hundreds to thousands of 4D CT images is necessary to evaluate airway collapsibility. Therefore, to enhance the clinical utility of 4D CT for the diagnosis of tracheobronchomalacia, it is imperative that an automatic or semiautomatic quantitative evaluation algorithm needs to be developed.

Similar to free-breathing cine CT (21), densitometric analysis of the lung parenchyma is feasible with free-breathing 4D CT. Consequently, we can easily differentiate a lung lesion with air trapping, from a normally ventilated lung region on 4D volume-rendered CT images (Fig. 9). It may be interesting to note that we may accentuate the visualization of lung ventilation, which is normally greater in the posterior dependent lungs, by freely adjusting the opacity of 4D CT images. Nevertheless, 4D CT-based lung ventilation imaging using a deformable image registration algorithm is desirable for the accurate and reproducible quantitative evaluation of regional ventilation function (22).

For diaphragm motion analysis, fluoroscopy, sonography, and magnetic resonance imaging (MRI) have all been used (1123). M-mode and B-mode sonography is most commonly used for diaphragmatic motion analysis in children, but the imaging method is limited by operator dependency and obscured acoustic window (24). A recent study (16) demonstrates that the diaphragmatic motion analysis and the densitometric analysis of adjacent lung parenchyma are feasible with free-breathing 4D CT, and the diagnosis of diaphragm paralysis or palsy is attainable in select children. In the literature, the normal diaphragm motion on M-mode sonography is defined as the respiratory excursion greater than 4 mm and the difference of the excursions between both sides less than 50%, irrespective of patient size and the degree of lung inflation (24). The lack of body size- and respiration-adapted normative data substantially hampers the accurate analysis of diaphragm motion in children. Hence, it is imperative to establish such normal data. In addition, a mobile CT unit capable of 4D CT scanning should be available to be used routinely in intensive care units.

To evaluate chest wall and aortic invasion by thoracic neoplasms pre-operatively, respiratory dynamic MRI has been used (2526). Dynamic 4D CT may also be used for this purpose. In the initial clinical experience in a small number of adult patients with lung cancer (18), 4D CT could reveal differential movements between the tumor and the adjacent structures in seven patients without direct invasion, as well as a small area of fixation due to superficial invasion of the adventitia of the superior vena cava in one case. There has been no such 4D CT study in children. Further clinical experiences are warranted to substantiate the clinical value of 4D CT in identifying pre-operative thoracic tumor invasion.

The most fundamental limitation in 4D thoracic CT evaluation is that normal data and diagnostic criteria are not established yet. Another potential limitation is that radiation dose of 4D CT may be increased not only for a longer scan range to include the entire thorax, but also for the increased number of the evaluated respiratory phases. Consequently, the radiation dose of 4D CT needs to be further reduced by utilizing a new reconstruction algorithm such as a penalized weighted least square smoothing method allowing reduced image noise and preserved spatial resolution (27). Because image degradation may occur at high respiratory rates and great respiratory excursions, a higher temporal resolution of 4D CT is desirable under the circumstances. Post-processing software dedicated for free-breathing 4D thoracic CT is vital to enhance its clinical utility. Compared to conventional CT scanning, the markedly increased demand for image storage and workload, proportional to the tremendously increased number of 4D CT images, should be considered. Relatively short longitudinal scan coverage, e.g., 3.8 cm in a 64-detector scanner, may be limited in evaluating the whole thorax with 4D CT imaging, but it is feasible in small children. When 4D CT scan is performed with intravenous injection of iodinated contrast agent, conscious sedation is usually required to maintain satisfactory image quality in young children. In contrast, sedation may be reduced or avoided in a non-contrast 4D CT scan.