Published online before print July 12, 2005, 10.1148/radiol.2363040306
(Radiology 2005;236:768-778.)
© RSNA, 2005
Obstructive Sleep Apnea in Pediatric Patients: Evaluation with Cine MR Sleep Studies1
Lane F. Donnelly, MD
1 From the Department of Radiology, Cincinnati Children's Hospital Medical Center, 3333 Burnet Ave, MLC 5031, Cincinnati, OH, 45229-3039. Received February 16, 2004; revision requested April 21; final revision received July 19; accepted August 4.
Address correspondence to the author (e-mail: Lane.Donnelly{at}cchmc.org).
 |
ABSTRACT
|
|---|
Cine magnetic resonance (MR) imaging sleep studies have become a useful tool in the evaluation of obstructive sleep apnea in children with certain categories of pathologic conditions. In this article, the author describes a program for the use of cine MR sleep studies in the evaluation of children with obstructive sleep apnea. The following areas are discussed: clinical indications, patient preparation, anatomic considerations, MR technique, technical issues, image interpretation, commonly encountered diagnoses, volume segmentation processing of data, and controversial areas.
© RSNA, 2005
|
INTRODUCTION
|
|---|
Obstructive sleep apnea is being increasingly recognized as a common disorder in children and can be associated with substantial morbidity. An association with obstructive sleep apnea has been identified for excessive daytime sleepiness, hyperactivity, attention deficit disorder, poor hearing, physical debilitation, and failure to thrive (1–6). It is estimated that up to 3% of all children are affected by obstructive sleep apnea (7,8).
A majority of children with obstructive sleep apnea are otherwise healthy children with enlarged adenoid and palatine tonsils. Evaluation of these children typically includes physical examination with direct inspection of the size of the palatine tonsils, as well as a lateral radiograph of the airway to evaluate the size of the adenoid tonsils and whether they encroach on the nasopharynx. A subgroup of patients with obstructive sleep apnea have more complex problems. This subgroup includes children with syndromes that predispose them to obstruction at multiple sites. Such conditions include Down syndrome, patients with craniofacial anomalies such as micrognathia, and children who have undergone failed prior surgery directed at the elimination of obstructive sleep apnea (9–13).
In such patients at our institution, magnetic resonance (MR) imaging with cine sequences has become a useful tool to help determine the anatomic and dynamic causes of persistent obstructive sleep apnea (9–13). Dynamic imaging studies, such as MR imaging or cine MR studies, have been shown to affect management decisions in over 50% of cases of complex obstructive sleep apnea (9–13). The advantage of cine MR sleep studies is that both static anatomy and dynamic abnormalities that lead to functional collapse of the airway are depicted (13–22).
In this review, we will describe our program for the use of cine MR sleep studies in the evaluation of children with obstructive sleep apnea. The following areas will be discussed: clinical indications, patient preparation, anatomic considerations, MR technique, technical issues, image interpretation, commonly encountered diagnoses, volume segmentation processing of data, and controversial areas.
|
CLINICAL INDICATIONS
|
|---|
All patients being considered for evaluation with cine MR imaging of obstructive sleep apnea are required to have documentation of obstructive sleep apnea based on the results of overnight polysomnography. Potential indications for cine MR evaluation of obstructive sleep apnea at our institution include persistent obstructive sleep apnea despite a previous surgical attempt to eliminate the obstructive sleep apnea (including previous tonsillectomy and adenoidectomy), obstructive sleep apnea with an underlying syndrome that predisposes the patient to potential multilevel obstruction, evaluation of supraglottic anatomy before anticipated complex airway surgery, and obstructive sleep apnea associated with obesity (13,21). In addition, patients with a tracheotomy tube who experience difficulty tolerating decannulation of the tracheotomy tube can also be evaluated with cine MR studies to determine if a supraglottic cause of the difficulty in decannulation is present.
Because of the involved nature of these studies, we highly recommend the establishment of defined indications and criteria for these studies. In addition, prior review of clinical information on referred patients is important. Occasionally, cine MR studies are inadvertently scheduled for patients who should more appropriately undergo other cross-sectional imaging studies of the airway, such as evaluation for extrinsic airway compression in a child with stridor.
|
PATIENT PREPARATION
|
|---|
Since symptoms occur only during sleep in children with obstructive sleep apnea, the children must be either asleep or sedated during cine MR evaluation. The decrease in muscular tone associated with natural sleep and sedation, as well as supine patient positioning, lead to changes in the anatomic positioning of structures that surround the airway and to increased dynamic motion of those structures, which lead to the findings of obstructive sleep apnea. Although it would be ideal to perform these studies during natural sleep, we have found this to be highly impractical and perform most if not all of our studies with the patient sedated. Attempts at the use of sleep deprivation to induce natural sleep during imaging is typically both inefficient and unsuccessful. The loud noise of the gradient-echo sequences used to create the cine MR images typically awakens subjects from natural sleep.
Until recently, the majority of cine MR studies were performed and monitored in accordance with our department's structured sedation program. A sedation nurse, pediatric radiologist, and respiratory therapist equipped with positive airway pressure breathing equipment were all present during sedation and the MR examination. The patient was sedated with intravenous injection of pentobarbital (3 mg per kilogram of body weight, with repeat doses if the patient remained awake, up to a total of 7 mg/kg) (13). During the entire procedure and sedation recovery, the patient's respiratory rate, heart rate and rhythm, and blood oxygen saturation levels were monitored by using transcutaneous pulse oximetry. If the patient had several obstructive sleep apnea episodes demonstrated during overnight polysomnography and if either the referring physician or radiologist monitoring the study was concerned about the prospect of sedation, the anesthesiology department was consulted to perform the sedation. Recent review of this process, as well as our previous experience with a radiology sedation program for sleep fluoroscopy in the evaluation of patients with obstructive sleep apnea, demonstrated the sedation program to be safe in such patients, with no complications related to sedation (10).
More recently, our institution has had increased involvement by the department of anesthesiology in the use of propofol for the sedation of patients undergoing an MR examination. Criteria for when an anesthesiologist performs propofol sedation for MR imaging have been expanded to include patients undergoing cine MR evaluation for obstructive sleep apnea. Now, nearly all patients evaluated for obstructive sleep apnea are sedated with propofol administered by an anesthesiologist. The decision to make this change in sedation policy was related, in part, to institutional changes and the increased involvement of the anesthesiology department in the MR suites, as well as to a concern that the method of determining which patients would be sedated according to the radiology sedation program and which would be referred to anesthesiology was subjective. Although the previous program had a perfect safety record, we think the presence of an anesthesiologist and sedation with propofol further decreases the potential risk of complications related to sedation.
|
ANATOMIC CONSIDERATIONS
|
|---|
As with all complex anatomic regions, there are various groupings of definitions and terms for the supraglottic area. Although which system of naming is chosen is not written in stone, an understanding of to what exactly the terminology refers is important for communication, particularly with regard to written radiology reports. The anatomic terminology used at Cincinnati Children's Hospital Medical Center is based on agreed-on terminology and definitions created by a group of radiologists, otolaryngologists, and pulmonologists who care for children with obstructive sleep apnea. The definitions differ somewhat from those that are often shown in traditional anatomy textbooks. The terms lend themselves to distinguishing anatomic areas where causes of obstructive sleep apnea commonly occur and are the anatomic definitions that have been used in numerous publications on imaging of patients with obstructive sleep apnea (9–16). These definitions will be used consistently throughout this article.
The definitions in use at our institution are as follows: The nasopharynx includes the aerated portion of the airway bordered by the soft palate anteriorly, the nasal turbinates anteriorly and superiorly, and the adenoids posteriorly and superiorly (Fig 1). The inferior border is at the level of the inferior tip of the uvula. The oral cavity is defined as the aerated space bordered by the hard palate superiorly, the tongue inferiorly, and the soft palate posteriorly (Fig 1). We define the hypopharynx as the aerated spaced bordered by the posterior aspect of tongue anteriorly, the posterior pharyngeal wall posteriorly, and the inferior aspect of the soft palate anteriorly (Fig 1). The inferior border is defined by the inferior extent (or base) of tongue (Fig 1). Note that in traditional anatomic descriptions, what we have designated as the hypopharynx is often referred to as the oropharynx. In a healthy sleeping child, the mouth is typically closed, the oral cavity is collapsed, and the nasopharynx and hypopharynx are patent with minimal wall motion (13–15,21) (Fig 1).
View larger version (137K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1a. Sagittal midline T1-weighted SE MR images (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) illustrate terminology and definitions for anatomy of the supraglottic airway. (a) Anatomy in a 14-year-old boy with cerebral palsy. Nasopharynx (light green) is defined as the aerated space bordered by soft palate anteriorly, adenoids posteriorly, and nasal turbinates anteriorly and superiorly. The inferior border is defined by the inferior tip of the uvula. Oropharynx (gray) is defined as the aerated space bordered by hard palate superiorly, tongue inferiorly, and soft palate posteriorly. Hypopharynx (dark green) is defined as the aerated spaced bordered by posterior aspect of tongue anteriorly, posterior pharyngeal wall posteriorly, and inferior aspect of soft palate anteriorly. The inferior border is defined by the inferior extent (or base) of the tongue. (b) More typical configuration encountered during normal sleep in a 9-year-old boy. The mouth is closed and the oral cavity (small arrowheads) is collapsed. The hypopharynx (arrows) and nasopharynx (large arrowhead) are patent.
|
|
View larger version (177K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1b. Sagittal midline T1-weighted SE MR images (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) illustrate terminology and definitions for anatomy of the supraglottic airway. (a) Anatomy in a 14-year-old boy with cerebral palsy. Nasopharynx (light green) is defined as the aerated space bordered by soft palate anteriorly, adenoids posteriorly, and nasal turbinates anteriorly and superiorly. The inferior border is defined by the inferior tip of the uvula. Oropharynx (gray) is defined as the aerated space bordered by hard palate superiorly, tongue inferiorly, and soft palate posteriorly. Hypopharynx (dark green) is defined as the aerated spaced bordered by posterior aspect of tongue anteriorly, posterior pharyngeal wall posteriorly, and inferior aspect of soft palate anteriorly. The inferior border is defined by the inferior extent (or base) of the tongue. (b) More typical configuration encountered during normal sleep in a 9-year-old boy. The mouth is closed and the oral cavity (small arrowheads) is collapsed. The hypopharynx (arrows) and nasopharynx (large arrowhead) are patent.
|
|
|
MR IMAGING TECHNIQUE
|
|---|
All cine MR studies are currently performed with 1.5-T MR units (Signa; GE Medical Systems, Milwaukee, Wis). With regard to the imaging field of view, the anatomic causes of obstructive sleep apnea involve the supraglottic airway, most commonly the nasopharynx and hypopharynx. Abnormalities that involve the subglottic airway, such as tracheomalacia or extrinsic tracheal compression, do not manifest as obstructive sleep apnea but as symptoms such as stridor. Therefore, it is essential that the supraglottic airway be imaged in its entirety, but it is not essential that the entire trachea be imaged in patients undergoing evaluation for obstructive sleep apnea.
Once asleep, the patient is placed in the head-and-neck vascular coil. With this coil, the airway from the superior aspect of the nasal passage to the level of the trachea can typically be imaged. In small patients, the entire airway can be visualized from its superior to its inferior extent (carina). In large adult-sized patients, the inferior aspect of the trachea may not be positioned in the imaging field of view. The patient is imaged with the cervical spine in neutral position. A three-dimensional localizer image is obtained. Sagittal and transverse T1-weighted spin-echo (SE) (repetition time msec/echo time msec of 400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) and transverse and sagittal fast SE inversion-recovery (IR) (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) images are obtained.
Cine MR sequences are performed in the midline sagittal location, as well as in the transverse plane, at the level of the middle portion of the tongue. The sequence used to create the cine MR images is a fast gradient-echo sequence (8200/3600, 80° flip angle, 12-mm section thickness). Approximately 128 consecutive images are obtained in the same location during an imaging time of approximately 2 minutes. Therefore, each image correlates with approximately 1 second. The sagittal or transverse images are then displayed in cine format and create a real-time "movie" of airway motion.
The midline sagittal plane is determined from a combination of the three-dimensional localization images and the sagittal T1-weighted images. The transverse plane at the level of the middle portion of the tongue (from superior to inferior) (Fig 2) is determined from the sagittal T1-weighted images.
View larger version (145K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2a. Optimal level to perform transverse cine MR imaging at level of the middle portion of the tongue is shown in a 14-year-old boy with cerebral palsy. (a) Sagittal midline T1-weighted SE image (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) shows appropriate level (line) for transverse cine images to be obtained. (b) Transverse gradient-echo cine MR image (8200/3600, 80° flip angle, 12-mm section thickness, 128 images obtained) obtained at level corresponding to the line in a. Arrows = hypopharynx, T = floor of mouth.
|
|
View larger version (128K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2b. Optimal level to perform transverse cine MR imaging at level of the middle portion of the tongue is shown in a 14-year-old boy with cerebral palsy. (a) Sagittal midline T1-weighted SE image (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) shows appropriate level (line) for transverse cine images to be obtained. (b) Transverse gradient-echo cine MR image (8200/3600, 80° flip angle, 12-mm section thickness, 128 images obtained) obtained at level corresponding to the line in a. Arrows = hypopharynx, T = floor of mouth.
|
|
Special Technical Considerations
The presence of dental braces should be identified and evaluated before sedation is induced in the patient. Dental braces are made of variable substances, and the degree of MR artifact caused by dental braces varies greatly. In many cases, the artifact caused by dental braces will obscure the region of interest within the nasopharynx and hypopharynx (Fig 3). In a child with dental braces who is able to cooperate for imaging without sedation, before sedation is induced we will place him or her in the head neck vascular coil and perform SE and cine imaging to determine the extent of MR artifact caused by the braces. If the artifact is minimal and does not interfere with the area of interest, sedation is induced and the remainder of the MR examination is performed. If artifact from the braces obscures the region of interest and the referring physician believes the MR evaluation is clinically essential, the patient is then rescheduled for MR imaging after removal of the dental braces (Fig 3).
View larger version (133K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3a. Artifacts associated with dental braces interfere with area of interest on sagittal midline T1-weighted SE MR images (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) in a 15-year-old boy. (a) With dental braces in place, marked artifact (arrows) obscures depiction of portions of area of interest. (b) After removal of dental braces, glossoptosis is seen. Posterior aspect of tongue (arrow) abuts posterior pharyngeal wall.
|
|
View larger version (138K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3b. Artifacts associated with dental braces interfere with area of interest on sagittal midline T1-weighted SE MR images (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) in a 15-year-old boy. (a) With dental braces in place, marked artifact (arrows) obscures depiction of portions of area of interest. (b) After removal of dental braces, glossoptosis is seen. Posterior aspect of tongue (arrow) abuts posterior pharyngeal wall.
|
|
Another special circumstance involves the patient who has obstructive sleep apnea that is so severe that the anesthesiologist has difficulty achieving a state in which the patient is adequately asleep and is able to maintain adequate oxygenation for the time required for the MR examination, without placement of an artificial airway. Such cases are the minority. We have adopted the following strategy in such patients: A nasal trumpet is placed, as the level of obstruction in these patients is typically either within the level of the nasopharynx or the hypopharynx at the middle portion of the tongue. The nasal trumpet bypasses the level of obstruction (Fig 4), and, in all such cases that we have encountered to date, the patient maintains adequate oxygenation with the nasal trumpet in place. Transverse and sagittal T1-weighted SE, fast SE IR, and gradient-echo cine images are then obtained with the nasal trumpet in place. After completion of these sequences, the nasal trumpet is removed, if possible, and sagittal cine images are obtained immediately (Fig 4). After this 2-minute sequence and if the patient is still maintaining adequate oxygenation, transverse cine images at the level of the middle portion of the tongue are acquired. In the rare event that the patient's oxygenation does not remain stable during imaging after removal of the nasal trumpet, the examination is terminated and the patient is immediately removed from the MR unit.
View larger version (150K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4a. MR imaging findings related to placement of nasal trumpet to establish patent airway. (a) Sagittal off-midline T1-weighted SE image (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) shows nasal trumpet (arrows) bypassing enlarged adenoids and glossoptosis in a 15-year old patient. Nasal trumpet passes inferior to level of the tongue. (b–e) Sagittal gradient-echo cine images (80° flip angle, 8200/3600, 12-mm section thickness, 128 images acquired) in 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy show changes in dynamic motion of the airway. (b, c) Images obtained at two points during respiratory cycle with nasal trumpet (not visible) show difference in diameter of hypopharynx (arrows) to be minimal. (d, e) Images obtained at two points during respiratory cycle without nasal trumpet in place show marked differences in diameter of hypopharynx (arrows), with intermittent collapse.
|
|
View larger version (135K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4b. MR imaging findings related to placement of nasal trumpet to establish patent airway. (a) Sagittal off-midline T1-weighted SE image (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) shows nasal trumpet (arrows) bypassing enlarged adenoids and glossoptosis in a 15-year old patient. Nasal trumpet passes inferior to level of the tongue. (b–e) Sagittal gradient-echo cine images (80° flip angle, 8200/3600, 12-mm section thickness, 128 images acquired) in 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy show changes in dynamic motion of the airway. (b, c) Images obtained at two points during respiratory cycle with nasal trumpet (not visible) show difference in diameter of hypopharynx (arrows) to be minimal. (d, e) Images obtained at two points during respiratory cycle without nasal trumpet in place show marked differences in diameter of hypopharynx (arrows), with intermittent collapse.
|
|
View larger version (130K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4c. MR imaging findings related to placement of nasal trumpet to establish patent airway. (a) Sagittal off-midline T1-weighted SE image (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) shows nasal trumpet (arrows) bypassing enlarged adenoids and glossoptosis in a 15-year old patient. Nasal trumpet passes inferior to level of the tongue. (b–e) Sagittal gradient-echo cine images (80° flip angle, 8200/3600, 12-mm section thickness, 128 images acquired) in 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy show changes in dynamic motion of the airway. (b, c) Images obtained at two points during respiratory cycle with nasal trumpet (not visible) show difference in diameter of hypopharynx (arrows) to be minimal. (d, e) Images obtained at two points during respiratory cycle without nasal trumpet in place show marked differences in diameter of hypopharynx (arrows), with intermittent collapse.
|
|
View larger version (155K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4d. MR imaging findings related to placement of nasal trumpet to establish patent airway. (a) Sagittal off-midline T1-weighted SE image (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) shows nasal trumpet (arrows) bypassing enlarged adenoids and glossoptosis in a 15-year old patient. Nasal trumpet passes inferior to level of the tongue. (b–e) Sagittal gradient-echo cine images (80° flip angle, 8200/3600, 12-mm section thickness, 128 images acquired) in 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy show changes in dynamic motion of the airway. (b, c) Images obtained at two points during respiratory cycle with nasal trumpet (not visible) show difference in diameter of hypopharynx (arrows) to be minimal. (d, e) Images obtained at two points during respiratory cycle without nasal trumpet in place show marked differences in diameter of hypopharynx (arrows), with intermittent collapse.
|
|
View larger version (147K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4e. MR imaging findings related to placement of nasal trumpet to establish patent airway. (a) Sagittal off-midline T1-weighted SE image (400/minimal, 22-cm field of view, 4-mm section thickness with 1-mm gap, 256 x 192 matrix, two signals acquired) shows nasal trumpet (arrows) bypassing enlarged adenoids and glossoptosis in a 15-year old patient. Nasal trumpet passes inferior to level of the tongue. (b–e) Sagittal gradient-echo cine images (80° flip angle, 8200/3600, 12-mm section thickness, 128 images acquired) in 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy show changes in dynamic motion of the airway. (b, c) Images obtained at two points during respiratory cycle with nasal trumpet (not visible) show difference in diameter of hypopharynx (arrows) to be minimal. (d, e) Images obtained at two points during respiratory cycle without nasal trumpet in place show marked differences in diameter of hypopharynx (arrows), with intermittent collapse.
|
|
Technical Issues
One issue that may be encountered is nonoptimal selection of the plane for the cine MR study in the midline sagittal plane, as well as selection of the transverse plane at the level of the middle portion of the tongue (Fig 2). In the event of suboptimal locations for these images, they should be reacquired. The sagittal image may be off midline. In our experience, when optimal positioning has not been achieved, the transverse image is often obtained too far inferiorly.
Another potential technical issue is related to the fact that the more severe the obstructive sleep apnea, the more the patient's head and mandible tends to "bob" (ie, move up and down quickly) during sleep. Therefore, the more severe the obstructive sleep apnea, the more motion artifact that can be identified on the MR images. In such cases, I will occasionally obtain the sagittal and transverse cine images on more than one occasion to increase the chance of acquiring optimal images.
With regard to positioning for other imaging studies, fast SE IR images are obtained mainly to evaluate the tonsillar tissue, which appears bright (Fig 5). Therefore, it is important for both the sagittal and transverse images to cover the entirety of the adenoids, as well as other tonsillar tissue. On occasion, transverse fast SE IR imaging will not be performed through the superior aspect of the adenoid tonsils. T1-weighted SE imaging in the transverse plane is the best sequence to perform for evaluation of intrathoracic tracheal pathologic conditions such as incidentally encountered tracheal extrinsic compression or tracheomalacia. Therefore, it is important to extend transverse T1-weighted imaging as inferiorly as possible, as dictated by the size of the patient and the length of the head-and-neck vascular coil. The fact that transverse T1-weighted SE and fast SE IR images do not cover the same anatomic distribution can be a source of confusion.
View larger version (172K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5a. Recurrent and enlarged adenoids in a 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy. (a) Sagittal midline fast SE IR MR image (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) shows recurrent and enlarged adenoids (A), which encroach on the nasopharynx. (b, c) Transverse gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images) obtained at level of base of the tongue. (b) At one point during respiratory cycle, airway (*) is patent. (c) At another point during respiratory cycle, airway shows cylindric collapse (arrows) with anterior, posterior, and lateral walls of hypopharynx having moved centrally.
|
|
View larger version (137K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5b. Recurrent and enlarged adenoids in a 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy. (a) Sagittal midline fast SE IR MR image (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) shows recurrent and enlarged adenoids (A), which encroach on the nasopharynx. (b, c) Transverse gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images) obtained at level of base of the tongue. (b) At one point during respiratory cycle, airway (*) is patent. (c) At another point during respiratory cycle, airway shows cylindric collapse (arrows) with anterior, posterior, and lateral walls of hypopharynx having moved centrally.
|
|
View larger version (136K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5c. Recurrent and enlarged adenoids in a 9-year-old girl with obstructive sleep apnea despite previous adenoidectomy. (a) Sagittal midline fast SE IR MR image (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) shows recurrent and enlarged adenoids (A), which encroach on the nasopharynx. (b, c) Transverse gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images) obtained at level of base of the tongue. (b) At one point during respiratory cycle, airway (*) is patent. (c) At another point during respiratory cycle, airway shows cylindric collapse (arrows) with anterior, posterior, and lateral walls of hypopharynx having moved centrally.
|
|
|
INTERPRETATION OF IMAGES
|
|---|
Cine images are used to evaluate for abnormal patterns of motion of the walls of the airway or, in other words, abnormal patterns of collapse of the airway. In a healthy sleeping child, the caliber of the airway in the nasopharynx and hypopharynx changes only minimally with respiration (13,14). Cine studies of healthy children demonstrate that there is rarely motion of the airway wall such that the luminal diameter of the hypopharynx or nasopharynx changes by more than 5 mm (13,14). Therefore, any motion greater than 5 mm should be considered abnormal. Certainly, intermittent collapse of the nasopharynx or hypopharynx is not encountered in patients without obstructive sleep apnea or other abnormalities (13,14). In a healthy sleeping child, the posterior nasopharynx (space between soft palate and adenoids) and the hypopharynx (space between posterior aspect of the tongue and posterior pharyngeal wall) should be patent. In contrast, it is not uncommon for the oral cavity between the superior aspect of the tongue and the hard and soft palates to be collapsed in a healthy sleeping child.
T1-weighted SE and fast SE IR images are used to evaluate the anatomy of the structures surrounding the airway. Fast SE IR images are particularly suited for evaluation of enlargement of the adenoid, palatine, and lingual tonsils. The tonsillar tissue has high signal intensity on fast SE IR images, compared with the remainder of the tissues, which have low signal intensity.
|
COMMON DIAGNOSES
|
|---|
Enlarged Adenoid Tonsils and Recurrent and Enlarged Adenoid Tonsils
The adenoids are absent at birth and then rapidly proliferate during infancy (15,23,24). The adenoids reach their maximum size between 2 and 10 years of age and then begin to progressively decrease in size beginning in the teenage years. There is some debate as to the upper limit of normal for maximum size of the adenoids. The maximum normal size of the adenoids is reported to be anywhere between 7 and 12 mm (15,23,24). Adenoids larger in size than 12 mm and associated with interim collapse of the posterior nasopharynx on the cine MR images should be considered enlarged (13,22) (Fig 5). The adenoids are measured on the midline sagittal image. The maximal diameter is measured at the level of the maximal convexity in the plane perpendicular to the anterior clival surface (23,24).
The relationship between enlargement of the adenoids and obstructive sleep apnea has been questioned (25–28). However, a previous study (15) in which a cine MR technique was used in asymptomatic patients demonstrated statistically significant relationships between increased size of the adenoids and both increased dynamic motion of the airway and increased prevalence of mouth breathing. These findings suggest that adenoid enlargement most likely does play a role in the development of obstruction in obstructive sleep apnea.
In most of the cases in which we perform an MR examination for obstructive sleep apnea, the patients have undergone tonsillectomy and adenoidectomy. In such patients, it is not uncommon to see some residual adenoid tissue as areas of high signal intensity in the region of the adenoid bed on fast SE IR images. The adenoid tonsils tend to be resected in more central than peripheral aspects of the adenoid fossa (Fig 6). Therefore, the residual amount of tissue tends to demonstrate a wedgelike central defect on transverse images. However, one of the most common causes of persistent or recurrent obstructive sleep apnea in patients who have undergone tonsillectomy and adenoidectomy is the presence of recurrent and enlarged adenoid tonsils (13,22). If the adenoids have recurred, are larger than 12 mm, and are associated with intermittent collapse of the posterior nasopharynx on sagittal cine images (Fig 5), the recurrent tissue should be considered a contributor to recurrent obstructive sleep apnea.
View larger version (166K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 6. Postoperative MR appearance of recurrent enlarged adenoid tonsils in 10-year-old boy with Down syndrome. Transverse fast SE IR image (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) shows enlarged adenoids (A), which measured 20 mm in anterior-posterior diameter. Note central wedgelike defect (arrow) in adenoids, which is related to previous adenoidectomy.
|
|
Enlarged Palatine Tonsils
Enlargement of the palatine tonsils is one of the most common anatomic causes of obstructive sleep apnea in the pediatric population. Enlarged palatine tonsils appear as round well-defined masses with high signal intensity on fast SE IR images and are positioned within the palatine tonsillar fossa (Fig 7). There are no publications of which we are aware regarding what size, as measured on images, should be considered normal or abnormal for palatine tonsils. However, if the palatine tonsils are seen as enlarged and are associated with inferior and central motion on cine images, leading to intermittent obstruction of the hypopharynx, they should be considered abnormally enlarged (13).
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7a. Fast SE IR MR images (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) of enlarged palatine tonsils in a 2-year-old boy with history of micrognathia and recurrent obstructive sleep apnea after mandibular extraction procedure. (a) Transverse image shows enlarged palatine tonsils (P) as well-defined high-signal-intensity structures in palatine fossa. (b) Sagittal image obtained through palatine fossa shows superior-to-inferior extent of enlarged palatine tonsil (P), as well as adenoids (A).
|
|
View larger version (161K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 7b. Fast SE IR MR images (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) of enlarged palatine tonsils in a 2-year-old boy with history of micrognathia and recurrent obstructive sleep apnea after mandibular extraction procedure. (a) Transverse image shows enlarged palatine tonsils (P) as well-defined high-signal-intensity structures in palatine fossa. (b) Sagittal image obtained through palatine fossa shows superior-to-inferior extent of enlarged palatine tonsil (P), as well as adenoids (A).
|
|
Since most of the patients who come to be evaluated with cine MR techniques for obstructive sleep apnea have already undergone removal of the palatine tonsils, we seldom see enlarged palatine tonsils on such studies. Unlike the adenoids, which commonly have recurred in patients with recurrent obstructive sleep apnea, it is rare in our experience that the palatine tonsils recur after surgical resection. Typically, in a patient who has undergone palatine tonsillectomy, no high-signal-intensity tissue is identified within the palatine tonsillar fossa on fast SE IR MR images.
Enlarged Lingual Tonsils
Enlargement of the lingual tonsils is a rare cause of obstructive sleep apnea, with only a handful of previous cases reported (22,29–33). Standards for the normal size range of the lingual tonsils as seen at imaging are not available. On fast SE IR MR images, normal lingual tonsils are seen as bilateral small crescentic areas of high signal intensity adjacent to the posteroinferior lateral aspects of the tongue. We have encountered, however, a number of cases in which enlargement of the lingual tonsils is a cause of obstructive sleep apnea. We have seen this most commonly in patients with Down syndrome who have undergone tonsillectomy and adenoidectomy (22). In such patients, the lingual tonsils are quite large (Fig 8). The tonsils appear as large masses arising from the base of the tongue and are high in signal intensity on fast SE IR images (22). In patients who have undergone palatine tonsillectomy, enlargement of the lingual tonsils can occasionally lead to growth superiorly into the palatine tonsillar fossa, as we have seen both at imaging and at surgery.
View larger version (158K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8a. Enlarged lingual tonsils in a 13-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) Sagittal and (b) transverse fast SE IR MR images (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) show enlarged lingual tonsil (arrows) filling and obstructing the hypopharynx.
|
|
View larger version (139K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 8b. Enlarged lingual tonsils in a 13-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) Sagittal and (b) transverse fast SE IR MR images (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) show enlarged lingual tonsil (arrows) filling and obstructing the hypopharynx.
|
|
Glossoptosis
Glossoptosis is defined as posterior motion of the tongue during sleep. In anatomic terms, the posterior aspect of the tongue moves posteriorly such that it intermittently comes in contact with the posterior wall of the hypopharynx, leading to obstruction of the airway at this level (10,11,13) ( Figs 3, 9). Glossoptosis can occur in relation to micrognathia (small mandible), macroglossia (large tongue), or decreased neuromuscular tone (10,11,13). Patients with Down syndrome are particularly predisposed to the development of glossoptosis because they have both macroglossia and decreased muscular tone. Patients with decreased muscular tone secondary to a neuromuscular disorder such as cerebral palsy are also at risk (11,13).
View larger version (175K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 9. Glossoptosis in 7-year-old boy with persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. Sagittal fast SE IR MR image (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) shows tongue positioned posteriorly, such that the posterior aspect of the tongue (arrow) abuts the posterior wall of the hypopharynx, leading to occlusion and displacement of the soft palate posteriorly (arrowheads). There is recurrence of the adenoids (A).
|
|
On sagittal cine MR images, the tongue and often the mandible intermittently move posteriorly and obstruct the hypopharynx (11,13) (Fig 10). In severe glossoptosis, the tongue can also push the soft palate posteriorly, causing intermittent obstruction of the nasopharynx. On transverse cine MR images obtained at the level of the middle portion of the tongue, the tongue is again seen to move posteriorly and intermittently obstruct the hypopharynx (Fig 11). Typically, the lateral aspects of the hypopharynx and the posterior wall of the hypopharynx remain stable in position, and the predominant motion is anterior-to-posterior motion of the tongue.
View larger version (165K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10a. Sagittal gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images) show glossoptosis in 8-year-old boy with persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) At one point during the respiratory cycle, the hypopharynx is patent, with the posterior aspect of the tongue (arrow) separated from the posterior wall of the hypopharynx. (b) At anther point during the respiratory cycle, the posterior aspect of the tongue has moved posteriorly, collapsing the hypopharynx (arrows) and displacing the soft palate posteriorly (arrowhead).
|
|
View larger version (166K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 10b. Sagittal gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images) show glossoptosis in 8-year-old boy with persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) At one point during the respiratory cycle, the hypopharynx is patent, with the posterior aspect of the tongue (arrow) separated from the posterior wall of the hypopharynx. (b) At anther point during the respiratory cycle, the posterior aspect of the tongue has moved posteriorly, collapsing the hypopharynx (arrows) and displacing the soft palate posteriorly (arrowhead).
|
|
View larger version (128K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11a. Comparison between hypopharyngeal collapse and glossoptosis with regard to hypopharyngeal wall motion patterns on transverse gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images). (a, b) Hypopharyngeal collapse in 8-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (b) Image obtained at same level as a but at another point during respiratory cycle shows cylindric collapse (arrows) of airway, with anterior, posterior, and lateral walls of hypopharynx moving centrally. (c, d) Glossoptosis in 13-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (c) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (d) Image obtained at same level as c but at another point during respiratory cycle shows airway collapse (arrows), which is related to posterior motion of the tongue rather than cylindric collapse of the airway.
|
|
View larger version (135K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11b. Comparison between hypopharyngeal collapse and glossoptosis with regard to hypopharyngeal wall motion patterns on transverse gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images). (a, b) Hypopharyngeal collapse in 8-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (b) Image obtained at same level as a but at another point during respiratory cycle shows cylindric collapse (arrows) of airway, with anterior, posterior, and lateral walls of hypopharynx moving centrally. (c, d) Glossoptosis in 13-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (c) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (d) Image obtained at same level as c but at another point during respiratory cycle shows airway collapse (arrows), which is related to posterior motion of the tongue rather than cylindric collapse of the airway.
|
|
View larger version (160K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11c. Comparison between hypopharyngeal collapse and glossoptosis with regard to hypopharyngeal wall motion patterns on transverse gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images). (a, b) Hypopharyngeal collapse in 8-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (b) Image obtained at same level as a but at another point during respiratory cycle shows cylindric collapse (arrows) of airway, with anterior, posterior, and lateral walls of hypopharynx moving centrally. (c, d) Glossoptosis in 13-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (c) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (d) Image obtained at same level as c but at another point during respiratory cycle shows airway collapse (arrows), which is related to posterior motion of the tongue rather than cylindric collapse of the airway.
|
|
View larger version (162K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 11d. Comparison between hypopharyngeal collapse and glossoptosis with regard to hypopharyngeal wall motion patterns on transverse gradient-echo cine MR images (8200/3600, 80° flip angle, 12-mm section thickness, 128 images). (a, b) Hypopharyngeal collapse in 8-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (a) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (b) Image obtained at same level as a but at another point during respiratory cycle shows cylindric collapse (arrows) of airway, with anterior, posterior, and lateral walls of hypopharynx moving centrally. (c, d) Glossoptosis in 13-year-old boy with Down syndrome and persistent obstructive sleep apnea despite previous tonsillectomy and adenoidectomy. (c) Image obtained at level of the base of the tongue at one point during respiratory cycle shows airway (*) to be patent. (d) Image obtained at same level as c but at another point during respiratory cycle shows airway collapse (arrows), which is related to posterior motion of the tongue rather than cylindric collapse of the airway.
|
|
Glossoptosis can be treated with surgical reduction of the posterior aspect of the tongue, somnoplasty (volume reduction of the posterior aspect of the tongue by means of submucosal cauterization of the tongue musculature), or anterior distraction of the mental portion of the mandible to pull the tongue anteriorly (21,34–42).
Hypopharyngeal Collapse
Hypopharyngeal collapse can occur as a primary phenomenon related to decreased muscular tone (increased elasticity of the hypopharyngeal walls) or as a secondary problem related to obstruction within a more superior part of the airway (9–13). A superiorly located obstruction of the airway, such as enlargement of the adenoid tonsils, creates increased negative pressure in the more inferior parts of airway during inspiration as an attempt to overcome the level of obstruction (Fig 5). This increased negative pressure can result in intermittent collapse of the hypopharynx (13). On cine MR images of hypopharyngeal collapse, there is intermittent cylindric narrowing of the hypopharynx (Fig 11). In other words, the posterior and left and right lateral walls of the hypopharynx and the posterior aspect of the tongue all move centrally. This is in contrast to the isolated posterior motion of the tongue seen in glossoptosis.
Transverse cine images at the level of the middle portion of the tongue are most helpful in differentiating glossoptosis from hypopharyngeal collapse (Fig 11). In contrast to the procedures for glossoptosis, surgical options for severe hypopharyngeal collapse include genioglossus advancement, hyoid myotomy, and hypopharyngeal suspension (21,35–42).
Abnormalities of the Soft Palate
An elongated soft palate is a reported contributing factor in the development of obstructive sleep apnea (40). Although there are no published guidelines regarding imaging criteria for when a soft palate should be considered abnormally elongated, criteria that we have found helpful include (a) a soft palate that is draped over and abutting the tongue and (b) the inferior aspect of the soft palate extending inferiorly below the middle portion of the tongue and associated with intermittent collapse of the nasopharynx on cine images. An elongated soft palate can be treated with uvulopalatopharyngoplasty (35–42).
At physical examination in patients with obstructive sleep apnea, the soft palate is often described as "woody" or edematous, findings that are thought to be related to the recurrent trauma of snoring. We have found that on sagittal fast SE IR MR images, the imaging correlate to the woody soft palate is increased signal intensity (unpublished observation) (Fig 12). The soft palate in patients without obstructive sleep apnea typically has low signal intensity, similar to that of the tongue musculature. We speculate that the increased signal intensity in the soft palate of some patients with obstructive sleep apnea may represent high water content due to edema. The presence of an edematous soft palate, seen as increased signal intensity of the soft palate as compared with that of the tongue on fast SE IR MR images, can be used as further confirmatory evidence of marked obstructive sleep apnea.
View larger version (171K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 12. MR findings of edematous soft palate in a 5-year-old girl with obstructive sleep apnea. Midline sagittal fast SE IR image (5000/34, echo train length of 12, 22-cm field of view, 6-mm section thickness with 2-mm gap, 256 x 192 matrix, two signals acquired) shows high signal intensity within soft palate (arrows), consistent with edema. Note high signal intensity of soft palate compared with that of tongue musculature (T).
|
|
|
COMMONLY ENCOUNTERED CLINICAL SCENARIOS
|
|---|
Down Syndrome
The management of refractory obstructive sleep apnea is particularly difficult in patients with Down syndrome because of the numerous potential causes of obstructive sleep apnea in these patients (22,35–40). Once a patient with Down syndrome has been shown to have persistent obstructive sleep apnea after tonsillectomy and adenoidectomy, our managing otolaryngologists often request cine MR studies to help direct further surgical intervention toward those supraglottic causes of sleep apnea that are demonstrated at imaging. Patients with Down syndrome are predisposed to macroglossia, glossoptosis, and enlarged palatine tonsils and adenoids. Also, patients with Down syndrome often have decreased muscular tone, which further exacerbates glossoptosis or hypopharyngeal collapse (22).
As previously mentioned in relation to the tendency for patients with Down syndrome to have lymphoid hypertrophy, these patients can experience marked hypertrophy of their lingual tonsils after tonsillectomy and adenoidectomy, which can explain persistent or recurrent obstructive sleep apnea (22). Adenoid regrowth is also a common cause of recurrent obstructive sleep apnea in these patients. In a review of patients with Down syndrome referred for cine MR evaluation (22), identified diagnoses included glossoptosis (63%), hypopharyngeal collapse (25%), recurrent and enlarged adenoids (63%), and enlarged lingual tonsils (17%).
Currently, there are no quantitative criteria for the diagnosis of macroglossia, and the imaging diagnosis of macroglossia is determined subjectively. The presence of macroglossia itself is not an important determination at imaging, because this is readily evaluated at a physical examination. What is more important is the determination of the presence and extent of associated glossoptosis. In addition to macroglossia, other abnormalities of the tongue in patients with Down syndrome include fatty infiltration, demonstrated as high signal intensity on T1-weighted MR images, and lack of a normal median sulcus (22).
Difficult Tracheotomy Tube Decannulation
Children with a tracheotomy tube are typically challenged before removal of the tube to see if they can tolerate breathing with the tracheotomy tube capped (occluded). Patients who have difficulty tolerating occlusion of the tracheotomy tube, particularly when this occurs during sleep, are often referred for MR evaluation for potential supraglottic causes of persistent airway obstruction. In such patients, when capping of the tracheotomy tube can be tolerated for the length of the MR imaging examination, the tracheotomy tube is capped and the MR protocol is performed as previously described. If there is difficulty in tolerance of capping of the tracheotomy tube, the entire series of sequences is obtained with the tracheotomy tube uncapped. If possible, the tracheotomy tube is then capped and the sagittal and transverse cine sequences are repeated. In patients undergoing imaging with both a capped and an uncapped tube, there typically is a greater degree of static collapse of the supraglottic airway and minimal motion of the airway walls when the tube is not capped. With the tracheotomy tube capped, there typically is a dramatic increase in the degree of airway motion (Fig 13). On all acquired images, the previously mentioned causes of obstructive sleep apnea are evaluated.
TOP
ABSTRACT
INTRODUCTION