The aim of this study was to assess the physiology of normal swallowing using recent advances in real-time magnetic resonance imaging (MRI). Therefore ten young healthy subjects underwent real-time MRI and flexible endoscopic evaluations of swallowing (FEES) with thickened pineapple juice as oral contrast bolus. MRI movies were recorded in sagittal, coronal, and axial orientations during successive swallows at about 25 frames per second. Intermeasurement variation was analyzed and comparisons between real-time MRI and FEES were performed. Twelve distinct swallowing events could be quantified by real-time MRI (start time, end time, and duration).
These included five valve functions: oro-velar opening, velo-pharyngeal closure, glottal closure, epiglottic retroflexion, and esophageal opening; three bolus transports: oro-velar transit, pharyngeal delay, pharyngeal transit; and four additional events: laryngeal ascent, laryngeal descent, vallecular, and piriform sinus filling and pharyngeal constriction. Repetitive measurements confirmed the general reliability of the MRI method with only two significant differences for the start times of the velo-pharyngeal closure (
Oropharyngeal dysphagia is a frequent sequelae caused by neuromuscular and neurological diseases or by structural and organic lesions of the oropharyngeal tract as well as in elderly patients [
At present videofluoroscopy reflects the gold standard in the diagnosis of deglutitive malfunctions. It offers dynamic images in the sagittal plane illustrating the complete course of deglutition but implies X-ray radiation exposure to patients. A coronal plane is usually also needed to exclude laryngeal penetration, whereas a view in the axial plane is not possible. Timing evaluation of the physiological events during swallowing relies exclusively on bony anatomic landmarks due to the limited visualization of soft tissues. On the other hand, flexible endoscopic evaluations of swallowing (FEES) offer an axial view and allow for pre- and postdeglutitive analyses that are valuable to clinical diagnosis and therapeutic planning in swallowing disorders. However, the “white out” at the moment of swallowing makes it difficult to reveal silent intradeglutitive aspiration and thus excludes the possibility for timing analyses. Recently, high-resolution manometry was reported to yield a robust predictor of aspiration in the diagnosis of oropharyngeal dysphagia [
Preliminary studies using magnetic resonance imaging (MRI) demonstrated potential for the evaluation of swallowing events and maneuvers in a supine position [
Ten healthy volunteers (4 men, 6 women) with a mean age of 28 ± 3 years (SD = standard deviation) and a range from 26 to 35 years were recruited from the local university. The selection criteria involved no history or presence of dyspnea, dysphonia, and dysphagia. They were fulfilled based on personal medical history and FEES examination of all subjects by an experienced otorhinolaryngologist (A. O.). The Institutional Review Board approved the study and all participants gave written informed consent prior to examination. MRI data of the same subjects have been used in a previous publication introducing the real-time technique [
Transnasal FEES was performed in a sitting position with a typical temporal resolution of 25 frames per second (fps). An oral bolus of one teaspoon (5 mL) green-colored pear pie was used to ensure a clear contrast to the tissues of the oropharyngeal tract. The flexible endoscope (Olympus ENF, Hamburg, Germany) was connected to a camera (Olympus visera OTVS-7, Hamburg, Germany) and the recorded videos were stored in a hard disk (rpSzene, Rehder & Partner GmbH, Hamburg, Germany) for further evaluation.
Dynamic MRI of deglutition in real time was performed with the use of a 3 Tesla MRI system (Tim Trio, Siemens Healthcare, Erlangen, Germany). The recently introduced real-time MRI technique [
Subjects were examined in a supine position with a combination of a small flexible coil (Siemens Healthcare, Erlangen, Germany) covering the lower face and a bilateral 2 × 4 array coil (NORAS MRI products, Hoechberg, Germany) centered to the thyroid prominence on both sides of the neck. Successive T1-weighted images (repetition time TR = 2.17 ms, echo time TE = 1.44 ms, flip angle 5°, field of view 192 × 192 mm2) were acquired with an in-plane resolution of 1.5 × 1.5 mm2 and a slice thickness of 10 mm in a midsagittal, oblique coronal, and oblique axial orientation. The total image acquisition time was 41.23 ms, which yielded a true temporal resolution of 24.3 fps without data interpolation or combination.
Pineapple juice was used as oral contrast agent due to its content of paramagnetic manganese, which leads to a bright signal in T1-weighted images [
To evaluate the swallowing events and their quantitative timings, the viewing software OsiriX (open-source software:
Absolute durations of each event were calculated by subtracting the timings of their end and start point images. For relative timings a reference event had to be chosen. Although the opening of the esophagus has been employed in previous studies [
Two-way repeated measures ANOVAs with the factors “measurements” and “events” were conducted to investigate possible differences for the start times, end times, and durations of defined events in two repetitive sagittal measurements. Post hoc paired
Real-time MRI movies were obtained from all subjects without any complications and complaints, while one FEES examination failed due to noncompliance of the subject to the transnasal procedure. Both MRI and FEES revealed normal deglutition and oromandibular function in all subjects. The use of a standardized real-time MRI examination protocol and a total in-room time of only about 15 minutes considerably reduced subject discomfort and facilitated the whole procedure.
A total of 12 distinct swallowing events were detected by real-time MRI. They are characterized as 5 valve functions, 3 bolus transports, and 4 additional events. A detailed description is given in Table
Swallowing events revealed by real-time MRI and videofluoroscopy.
Deglutition events | Real-time MRI | Videofluoroscopy [ |
---|---|---|
Valve function | ||
Oro-velar opening (OOT*, reference) | Full rise of soft palate from dorsum | Not defined |
Velar-pharyngeal closure (VCT) | Contact of soft palate and pharyngeal wall | Velopharyngeal sphincter (valve 3) |
Glottal closure (GCT) | Closure of glottis and supraglottis | Larynx: vocal folds (valve 4c) |
Epiglottic retroflexion (ERT) | Contact of epiglottis and supraglottis | Larynx: epiglottis and arytenoid to base of epiglottis (valve 4a/b) |
Esophageal opening (EOT) | Separation of postcricoid and pharyngeal wall | Cricopharyngeal sphincter (valve 6) |
Bolus transport | ||
Oro-velar transit (OTT) | Oral transit | |
s: OOTs | s: initiation of tongue movement | |
e: OOTe | e: bolus head reaches the cross of mandible and tongue base | |
Pharyngeal delay (PDT) | s: OTTs (OOTs) | s: bolus head reaches cross of mandible and tongue base |
e: PTTs | e: start of laryngeal ascent | |
Pharyngeal transit (PTT) | s: bolus head passes oro-velar valve | s: start of laryngeal ascent |
e: bolus tail passes esophageal sphincter | e: bolus tail passes cricopharyngeal region | |
Other | ||
Laryngeal ascent (LAT) | Ascendance of larynx | Upward and forward movement of hyoid and larynx |
Laryngeal descent (LDT) | Descendance of larynx | Not defined |
Vallecula and piriform sinus filling (SFT) | Bolus filling vallecula and piriform sinus | Not observed |
Pharyngeal constriction (PCT) | Progressive contraction of pharyngeal constrictor | Tongue base and pharyngeal wall (valve 5). Not observed [ |
Physiological events of normal swallowing as seen by real-time MRI (27-year-old female). LAT: laryngeal ascent, VCT: velo-pharyngeal closure, OOT: oro-velar opening (start time defined as reference), PTT: pharyngeal transit, GCT: glottal closure, SFT: vallecular and piriform sinus filling, ERT: epiglottic retroflexion, PCT: pharyngeal constriction, EOT: esophageal opening, LDT: laryngeal descent (“s” and “e” refer to respective start and end times). The images are selected from respective movies (see Supplementary Movie 1 in the Supplementary Material available online at
The three events that describe the passing of the bolus rather than the behavior of the valve refer to the oro-velar transit (OTT), pharyngeal delay (PDT), and pharyngeal transit (PTT). The OTT represents the bolus transport through the oro-velar valve, whose start and end were coincident with that of the OOT. The PTT was defined as the duration from the onset of the bolus head passing the oro-velar valve (start) to the point where the bolus tail passes the esophageal sphincter (end), while the PDT was the interval between the OOT start and the onset of the bolus head passing the oro-velar valve. Additional events included the laryngeal ascent (LAT), laryngeal descent (LDT), vallecular and piriform sinus filling (SFT), and pharyngeal constriction (PCT). The latter two events were again newly detected by real-time MRI.
Table
Detectability of swallowing events by real-time MRI and FEES.
Deglutition events | Real-time MRI | FEES | |||
---|---|---|---|---|---|
Sagittal | Coronal | Axial | Axial | ||
Laryngeal ascent (LAT) | Start |
|
80% | 0% | 89% |
End |
|
67% | 0% | ||
Velo-pharyngeal closure (VCT) | Start |
|
|
5% |
|
End |
|
|
5% |
|
|
Oro-velar opening (OOT) | Start |
|
53% | 0% | 33% |
End |
|
7% | 0% | ||
Pharyngeal transit (PTT) | Start |
|
93% | 0% | |
End |
|
20% | 0% | ||
Glottal closure (GCT) | Start | 0% |
|
|
|
End | 0% |
|
|
22% | |
Vallecular and piriform sinus filling (SFT) | Start | 11% | 67% |
|
|
End | 0% | 33% |
|
||
Epiglottic retroflexion (ERT) | Start |
|
0% | 5% | |
End |
|
0% | 0% | 89% | |
Pharyngeal constriction (PCT) | Start |
|
|
5% | |
End |
|
|
5% | ||
Esophageal opening (EOT) | Start |
|
33% | 68% | |
End |
|
13% | 68% | ||
Laryngeal descent (LDT) | Start |
|
47% | 0% | |
End |
|
53% | 0% | 11% |
Bold font: highest visibility rate.
Bold italic font: second highest visibility rate.
Figure
Durations, start times, and end times of distinct swallowing events as determined by real-time MRI (means and quartiles, 10 subjects). LAT: laryngeal ascent, VCT: velo-pharyngeal closure, OOT: oro-velar opening (start time defined as reference), OTT: oro-velar transit, PDT: pharyngeal delay, PTT: pharyngeal transit, GCT: glottal closure, SFT: vallecular and piriform sinus filling, ERT: epiglottic retroflexion, PCT: pharyngeal constriction, EOT: esophageal opening, and LDT: laryngeal descent.
Data from the two repetitive sagittal measurements were compared with ANOVAs for durations (Figure
Durations of distinct swallowing events for two repetitive real-time MRI measurements (means and quartiles, 10 subjects). LAT: laryngeal ascent, VCT: velo-pharyngeal closure, OOT: oro-velar opening (start time defined as reference), OTT: oro-velar transit, PTT: pharyngeal transit, GCT: glottal closure, SFT: vallecular and piriform sinus filling, ERT: epiglottic retroflexion, PCT: pharyngeal constriction, EOT: esophageal opening, and LDT: laryngeal descent.
Start and end times of laryngeal ascent (LAT) and descent (LDT) for two repetitive real-time MRI measurements (means and quartiles, 10 subjects).
The duration of the velo-pharyngeal closure was compared between real-time MRI and FEES, as it is the only event detectable by the latter. Sagittal and coronal real-time MRI studies resulted in 724 ± 144 ms and 690 ± 112 ms, respectively, which strongly correlate (
This study demonstrates that recent advances in real-time MRI [
When compared to Logemann’s model [
For bolus transport events, the establishment of a landmark location appears to be a major challenge. So far, there is no generally accepted standard. Previous definitions refer to the “bolus head reaching the cross of the tongue base and posterior aspect of the mandible ramus” and “start of laryngeal elevation” to distinguish between oral transit, pharyngeal delay, and pharyngeal transit [
Similar concerns apply to the laryngeal ascent, which was previously defined as a landmark for the pharyngeal transit and its timings (PTT). The present data show that the laryngeal ascent is one of only two events with a significant difference in the start time between two repeated measurements, indicating that various laryngeal movements are involved at the early phase of swallowing [
Based on these definitions, a delay ranging from zero to 200 ms (78 ± 66 ms) was observed between the wide opening of the oro-velar valve (OOTs) with the bolus ready to enter the pharyngeal cavity and the onset of the bolus head passing the oro-velar valve (PTTs)—here defined as pharyngeal delay (PDT). Although using a different landmark, its timing is in accordance with earlier videofluoroscopy findings [
Despite the limited number of subjects, our results clearly demonstrate that the sagittal plane is the most valuable orientation for the assessment of most physiological swallowing events. This also applies to conventional videofluoroscopy and may further simplify clinical applications by restricting the MRI examination to the sagittal plane at even shorter scan times. For example, this would particularly be favorable for patients suffering from dysphagia.
In the analysis of individual swallowing events, the choice of a landmark or reference is of great importance. So far, the beginning of the esophageal opening (EOTs) was taken as the reference [
Here the descriptive statistics has revealed different levels of variance associated with individual swallowing events across the subjects. In particular, LAT, VCT, and LDT show larger variation in durations whereas all the other events have no difference in relative timings and durations from one measurement to the other. This is better illustrated in Figure
Taken together, the present real-time MRI findings unravel a well-orchestrated temporal pattern of physiological swallowing events during deglutition, as summarized in Figure
Temporal pattern of physiological events during normal swallowing. VCT: velo-pharyngeal closure, ERT: epiglottic retroflexion, GCT: glottal closure, EOT: esophageal opening, OOT: oro-velar opening (start time defined as reference), OTT: oro-velar transit, PDT: pharyngeal delay, and PTT: pharyngeal transit. Solid lines indicate durations with mean values in brackets (10 subjects), while circles represent respective start (solid) and end times (open). Arrows refer to maximal durations.
A major difference between real-time MRI and FEES is the upright body position for FEES and the supine position for MRI. Because statistics based on VCT duration revealed a significant difference between both methods, it seems that the supine position may have some impact on the exact time course of swallowing. This finding needs further investigation of systematic impact of the body position on deglutitive timings. However, any direct comparison will be compromised by the limited number of visible events in FEES and the technical restrictions precluding a simultaneous dual-modality examination. On the other hand, patient studies of swallowing in the supine position have been reported [
In conclusion, the results of this preliminary real-time MRI study at 41 ms temporal resolution offer comprehensive information about the physiology of normal swallowing and the function of the dynamic events. Real-time MRI contributes to our understanding of swallowing by providing images in arbitrary orientations that cover the entire oropharyngolaryngeal region. It also allows for the establishment of new landmarks and standards and provides access to a detailed description of the temporospatial pattern of the swallowing process. The present reference values for normal swallowing in the supine position will serve as the basis for further investigations of pathological conditions as in an ongoing study of dysphagia patients. Other real-time MRI applications may address therapeutic interventions such as functional swallowing therapies. In general, clinical studies are now warranted to assess the real-time MRI potential for diagnosis and treatment monitoring of swallowing disorders.
Eesophageal opening
Epiglottic retroflexion
Flexible endoscopic evaluations of swallowing
Fast low-angle shot
Glottal closure
Laryngeal ascent
Laryngeal descent
Magnetic resonance imaging
Oro-velar opening
Oro-velar transit
Pharyngeal constriction
Pharyngeal delay
Pharyngeal transit
Vallecular and piriform sinus filling
Velo-pharyngeal closure.
The authors declare that there is no conflict of interests regarding the publication of this paper.
The first two authors have contributed equally to this study. Arno Olthoff and Shuo Zhang designed the study, performed the examination and the analysis, interpreted the results, and drafted the paper. Renate Schweizer performed the statistical analysis and helped to interpret the results and to draft the paper. Jens Frahm conceived and supervised the study, interpreted the results, and revised the paper. All authors read and approved the final paper.
The authors thank Dr. Kai Kallenberg, Dr. Marios Psychogios, Dr. Tibor Auer, and Meike Schweisfurth for valuable discussions. The authors also thank Sebastian Schaetz for online image reconstructions using a dedicated algorithm and GPU-based computer fully integrated into the software architecture of the used MRI system.