| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Pediatric Imaging |
1 From the Departments of Radiology (M.B.A., L.F.D., B.J.D.) and Pediatrics (L.F.D., B.J.D., S.A.P., B.A.C., R.S.A.) and Division of Pulmonology (B.A.C., R.S.A.), Cincinnati Children’s Hospital Medical Center, 3333 Burnet Ave, Cincinnati, OH 45229-3090. Received October 1, 2003; revision requested December 23; revision received January 7, 2004; accepted February 2. Address correspondence to L.F.D. (e-mail: lane.donnelly@cchmc.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Transverse fast gradient-echo cine MR images of the hypopharynx were obtained at 1.5 T in 31 children with OSA (eight girls, 23 boys; mean age, 11.3 years) and 21 children free of airway symptoms who underwent MR imaging for other clinical indications (11 girls, 10 boys; mean age, 3.5 years). Volume segmentation with a k-means clustering algorithm was applied to transverse cine MR images to quantify airway volumes at each time. Airway wall motion for each child was described with standard deviation and range. Coefficient of variance and normalized range, which are independent of airway size, were used to compare groups (Kruskal-Wallis test).
RESULTS: Plots of airway volume over time demonstrated large fluctuations during respiration in children with OSA and minimal fluctuations in controls; findings were consistent with airway distention and airway collapse in OSA. Average airway transverse volume was larger in the group with OSA than in the control group (OSA group, 2.52 mL; control group, 0.936 mL; P < .001). Mean standard deviation (OSA group, 0.840 mL; control group, 0.17 mL; P < .001) and mean range of airway cross section (OSA group, 3.552 mL; control group, 0.864 mL; P < .001) were larger in the group with OSA. Coefficient of variance (OSA group, 0.32; control group, 0.17; P < .001) and normalized range (OSA group, 1.42; control group, 0.96; P < .001) indicate statistically significant difference in airway dynamics in children with OSA.
CONCLUSION: Volume segmentation of transverse cine MR images of the hypopharynx aids in quantification of increased airway wall motion in children with OSA. Transverse MR imaging demonstrates both airway distention and collapse in children with OSA.
© RSNA, 2004
Index terms: Magnetic resonance (MR), cine study, 27.121412, 27.121419 • Magnetic resonance (MR), in infants and children, 27.121412, 27.121419 • Sleep apnea, 27.827
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Imaging has been useful in diagnosis and characterization of OSA. Historically, imaging has almost exclusively included static images in the lateral or sagittal view. The lateral view is most useful in sleep fluoroscopy and provides an excellent view of the anterior to posterior dimensions and movements along the length of the airway (9,10,14,15). Similarly, cephalometrics includes standard measurements of structures in the head and neck, and the lateral view is the most useful for visualization of the airway, with a minimum of superimposed structures (16–19). Cross-sectional imaging available from both computed tomography (CT) and magnetic resonance (MR) imaging has provided improved anatomic resolution (20–25). Recently, cine MR imaging has combined the benefits of cross-sectional imaging with visualization of airway dynamics (26–28).
Cine MR imaging can be applied in multiple planes, with the most useful typically being the sagittal and transverse planes. Sagittal cine MR images are reminiscent of sleep fluoroscopic studies and provide exquisite views of the airway from nares to trachea (20,26–28). The use of transverse MR imaging in OSA is relatively unexplored (28,29). Transverse cross-sectional images of the airway complement the sagittal views. Since motions of the airway in the lateral dimension can be observed, transverse imaging may be useful in distinguishing between mechanisms of airway obstruction, such as glossoptosis and hypopharyngeal collapse. In a study (28) of 16 children with OSA and 16 control subjects imaged with both transverse and sagittal cine MR imaging techniques, investigators identified the hypopharynx as the most common site of airway collapse. In that study, transverse cine MR images demonstrated substantial airway collapse in the left-to-right dimension, rather than in the anterior-to-posterior dimension, in children with OSA—a finding not observed on the sagittal view. The purpose of our study was to retrospectively determine airway wall motion with volume segmentation of transverse cine MR images in children with OSA.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Control subjects were children without respiratory symptoms who were imaged for other clinical indications with sedation. With institutional review board approval, informed consent was obtained from the parents to perform the cine MR imaging sequence. The cine MR imaging sequence was performed following completion of the clinically indicated MR imaging sequences in children who were undergoing MR imaging of the brain and required sedation in order to be able to cooperate with personnel during the examination. We excluded from the control group those children who had histories, signs, or symptoms of airway abnormalities (previous airway surgery, tracheotomy, OSA, snoring) detected at the presedation work-up. Also, we excluded those children who experienced oxygen desaturation or noisy breathing during sedation. Since the control group consisted of children who required sedation to be able to cooperate with personnel during the MR imaging examination, most of these children were younger than 6 years old. Data from the control group were obtained during a shorter time than that in which the data for the OSA group were obtained. Age and sex of those in the control and OSA groups were recorded.
Imaging
Children in the OSA group were sedated for imaging to simulate sleep, according to our clinical protocol for MR imaging sleep studies. Sedation in both the OSA and control groups was induced according to our institutional sedation protocol, with the supervision of a pediatric radiologist, a pediatric nurse, and a respiratory therapist (30,31). Oral or intravenous sedatives were administered to the children per our in-house protocol. Chloral hydrate, 50–100 mg per kilogram body weight, was administered to children younger than 1 year. Pentobarbital, 3 mg/kg (with additional doses up to a total of 7 mg/kg), was administered to children 1 year or older. There were no sedation-related complications. Among the control subjects, images for the current study were obtained following performance of the clinically indicated studies. In no case was additional sedation induced to complete an experimental sequence. If a child awakened during acquisition of the experimental study images, the sequence was aborted.
In the OSA group, the children were placed supine with the neck in a neutral position in the head-neck vascular coil. In the control group, subjects were imaged in a head coil following completion of clinically requested MR imaging of the brain. Among the various MR imaging series conducted in these subjects, cine MR images were obtained with a 1.5-T MR imager (Signa; GE Medical Systems, Milwaukee, Wis) and a fast gradient-echo pulse sequence (repetition time msec/echo time msec, 8.2/3.6; flip angle, 80°; section thickness, 12 mm; field of view, 24 cm; in-plane resolution, 256 x 256). Cine MR images were obtained in the transverse plane at the level of the middle of the tongue. Images were obtained at the same level of the hypopharynx in all subjects. The appropriate level was chosen on the basis of data from a three-dimensional localizer and of information about other imaging sequences previously performed, most commonly sagittal T1-weighted spin-echo MR imaging. One hundred twenty-eight consecutive images were obtained, with a total imaging time of approximately 2 minutes. These images were displayed in cine format to create a movie of airway motion for qualitative evaluation. All clinical studies were reviewed at the time they were obtained by a pediatric radiologist (L.F.D.).
Volume Segmentation Analysis
For volume segmentation analysis, the cine MR images in the Digital Imaging and Communications in Medicine file format were transferred to a computer workstation. Volume segmentation of the cine MR images was implemented with an in-house image analysis software program called CCHIPS, or Cincinnati Children’s Hospital Image Processing Software (Research Systems, Boulder, Colo). Within CCHIPS, a single matrix of signal intensity data that encompassed the entire cine time course was constructed from the 128 separate MR images. The region of interest for volume segmentation was manually selected by a single reviewer (M.B.A.) to encompass the entire cross section of the hypopharynx. The matrix of signal intensity data within the region of interest was then analyzed with the k-means clustering segmentation algorithm within CCHIPS (29). The region of interest was segmented for the entire data set (all sections over all time evaluated) into three signal intensity levels. Three segments yielded an accurate depiction of airway borders that was based on a comparison with clinical images.
The segmented signal intensity data were displayed by assigning each signal intensity level a particular shade of gray. The two lighter shades represented the signal intensity of tissues adjacent to the airway, and the darkest gray area represented the signal intensity of the airway itself. Two soft-tissue segments were used to compensate for motion artifacts and intrinsic air–soft-tissue interface artifacts. By using a third segment to account for noise, the volumes calculated were thought to more accurately reflect the patent portion of the airway (29). The image segment that represented the patent airway was selected, and airway volume was plotted as a function of time. Segmentation of two-dimensional images allows quantification of a volume, since the imaged planes have an intrinsic thickness. Since the imaged plane is of near uniform thickness, the segmentation volumes are directly proportional to the cross-sectional area of the imaged airway.
Data and Statistical Analysis
Change in volume over time was normalized for average airway volume for the subject and plotted across the time course. The average airway volume was calculated by using the mean airway volume measured during the entire period that the cine imaging sequence was performed. Normalization allows the change in volume of the area to be viewed in relationship to the relative size of the subject’s airway, as this may vary according to size and age of subjects and plane of imaging. The changes in the normalized volume over time can then be demonstrated graphically. Volume segmentation allows objective and precise quantification of airway size and can be automated to analyze large numbers of images in a consistent manner. Volume segmentation was applied to the transverse cine MR images to quantify changes in airway size over time. Numeric characterization of airway size allowed statistical analysis of airway wall motion and objective comparison of children with OSA and control subjects.
The standard deviation (SD) and range of airway volumes were calculated for each child. These parameters are measures of individual airway wall motion and were also normalized for airway volume to allow comparison of individuals of different sizes.
Statistical comparisons of average volume, SD, range, normalized SD (coefficient of variance), and normalized range between children with OSA and control subjects were made by using the Kruskal-Wallis test. The Kruskal-Wallis test was used because the data are not normally distributed. The effects of age on airway dynamics parameters between groups were compared with the two-sample t test. The effects of sex on airway dynamics parameters between groups were compared with the 
2 test. Values with P < .05 were considered to indicate a statistically significant difference.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
|
|
|
|
|
Average airway transverse volume was larger in the OSA group than in the control group (OSA group, 2.52 mL; control group, 0.94 mL; P < .001). There is the potential for this value to be confounded by age, as the control group included younger subjects (mean age, 3.5 years; SD, 2.17) than did the OSA group (mean age, 11.3 years; SD, 6.05). To investigate the contribution of age to airway size, the average airway volume was plotted according to the subject’s age (Fig 5). Some of the youngest children with OSA had airway volumes as large as some of the oldest control subjects. Nearly horizontal trend lines (least squares fit) graphically depict the poor correlation of the airway volume with the subject’s age. Correlation coefficients of R2 = 0.027 for children with OSA and of R2 = 0.144 for control subjects suggest little dependence of airway volume on age within each group. There are subjects between the ages of 40 and 100 months in both groups. Figure 5 demonstrates that these subsets of subjects have different average airway volumes. Airway volume did not correlate with age within either study group, which supports the conclusion that a substantial portion of the difference in airway volume between groups is associated with OSA.
|
In the OSA group, there were eight girls and 23 boys. In the control group, there were 11 girls and 10 boys. There was no statistically significant difference in the same imaging parameters on the basis of sex (all P > .05).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Airway distention secondary to flow obstruction may be explained by consideration of fluid dynamics. An upstream narrowing of the airway that causes resistance to airflow will result in greater negative pressure in the distal airway. The greater negative pressure may cause the soft-tissue walls of the hypopharynx to collapse during inspiration. Conversely, the same proximal airway narrowing will provide resistance to airflow during expiration. This resistance creates positive pressure that causes airway distention during expiration. Airway collapse and expansion during respiration would be exaggerated in the setting of increased airway elasticity or compliance.
Airway obstruction is thought to be the defining phenomenon in OSA. OSA is known to result from anatomic obstructions, which include enlarged tonsils, macroglossia, micrognathia, and retrognathia (9–13). Airway compliance or elasticity, however, has also been implicated in OSA (32–38). Mechanic and mathematic models of airway dynamics consistently implicate airway elasticity in airway dynamics and collapse (33,37,39,40). Increased airway tissue elasticity in adults with OSA has been demonstrated with a variety of in vivo and ex vivo methods (32,34,35,41,42). Hypotonia and increased elasticity are associated with OSA in children in the absence of lymphoid tissue hyperplasia, with certain syndromes such as Down syndrome, and with other anatomic causes of obstruction (43). Though theories about the cause that incorporate both anatomic obstruction and airway compliance have been proposed (44), the relative importance of anatomic obstruction and airway compliance has been intensely debated (45,46).
The ballooning and collapse of the upper airway during expiration in children with OSA in this study may support a role for airway compliance in the pathogenesis of OSA. The theory of a simple anatomic cause of OSA may not predict and completely account for the degree of airway distention observed in this study. Future studies with simultaneous measurement of inspiratory effort and airflow and correlation of these findings with findings at imaging of the airway may provide more insight into the pathophysiology of OSA.
This study had limitations. We recognize that there was an age discrepancy between the OSA group and the control group. This discrepancy was caused by the characteristics of the subjects included in the control group: This group included subjects who were being imaged for other clinical indications with sedation. Children who need sedation to cooperate with personnel to undergo MR imaging tend to be younger than 5 years. OSA is found in children of all ages, but typically cine MR imaging is performed in children older than 5 years. This is why ages of subjects in the control and study groups were different. Since it would be unethical to expose control subjects to the risks of sedation for experimental purposes, an age-matched control study of children with OSA is unlikely to occur. The primary concern is that airway size may increase with age. Interestingly, some of the youngest subjects with OSA had some of the largest airways. Airway volume did not correlate significantly with age in either the control group or the children with OSA. Measures of airway dynamics were normalized to account for size differences between subjects. Differences in airway dynamics between groups remained statistically significant after normalization of the data to allow for differences in airway size between subjects. Therefore, we have done our best to deemphasize the size issues related to age between the control and OSA groups. It is also possible that subjects of different ages have different airway physiology and elasticity. This study does not address these issues.
The second limitation was that in this study subjects in both the control and OSA groups were sedated. Some argue that sedation does not completely mimic physiologic sleep. Although this is true, during both sedation and physiologic sleep, the muscular tone of the airway is decreased and the two states may be similar. For two decades, physicians at our institution have been using the results of imaging performed while children were sedated to make clinical decisions concerning those with problematic OSA. In addition, children with OSA tend to have noisy breathing and oxygen desaturation when sedated with these techniques in patterns that the parents report are similar to those that develop during sleep at home. In contrast, it is uncommon for children without OSA to demonstrate such findings when they are sedated. These findings also support the notion that sedation is similar to physiologic sleep. In this study, since subjects in both the OSA and control groups were sedated in an identical fashion, the statistically significant differences in airway motion can be attributed to OSA. In addition, we found that attempts to perform cine MR imaging in subjects during physiologic sleep are not practical and usually fail; in part, this failure is related to the loud sound generated by the sequences used to create the cine MR images.
A final limitation of this study was that since the children in the OSA group had anatomic causes of OSA, these anatomic anomalies had an effect on the degree of airway motion. It is impossible to completely separate the changes in airway wall motion that occur as a result of anatomic narrowing from those changes in airway wall motion that result from issues related to wall elasticity.
In conclusion, airway wall motion, which was quantitatively identified in children with OSA, was increased in comparison with that in control subjects. Substantial airway distention in children with OSA may further implicate airway tissue elasticity in the pathophysiology of OSA.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Abbreviations: OSA = obstructive sleep apnea, SD = standard deviation
Author contributions: Guarantor of integrity of entire study, L.F.D.; study concepts and design, M.B.A., L.F.D.; literature research, M.B.A., L.F.D.; clinical studies, all authors; data acquisition, all authors; data analysis/interpretation, M.B.A., L.F.D.; statistical analysis, S.A.P.; manuscript preparation, M.B.A., L.F.D.; manuscript definition of intellectual content and editing, all authors; manuscript revision/review and final version approval, L.F.D.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  L. F. Donnelly Obstructive Sleep Apnea in Pediatric Patients: Evaluation with Cine MR Sleep Studies Radiology, September 1, 2005; 236(3): 768 - 778. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) and control subjects (
). Nearly horizontal trend lines (least squares fit) graphically depict the poor correlation of airway volume with subject age. Correlation coefficients of R2 = 0.027 for subjects with OSA (dashed line) and R2 = 0.144 for control subjects (solid line) suggest little dependence of airway volume on age within each group. There were a substantial number of subjects between the ages of 40 and 100 months in both groups. These subsets of subjects had different average airway volumes. Airway volume did not appear to be correlated with age within either control or OSA group, which supports the conclusion that a significant portion of the difference in airway volume between groups was associated with OSA rather than with age.