1Department of Physical Medicine and Rehabilitation, National Taiwan University Hospital Hsin-Chu Branch, Hsin-Chu, Taiwan
2Department of Physical Medicine and Rehabilitation, National Taiwan University Hospital, Taipei City, Taiwan
3Department of Physical Medicine and Rehabilitation, College of Medicine, National Taiwan University, Taipei City, Taiwan
Correspondence: Ming-Yen Hsiao Department of Physical Medicine and Rehabilitation, College of Medicine, National Taiwan University, No.7, Zhongshan S. Rd., Zhongzheng Dist., Taipei City 100, Taiwan. Tel: +886-972652857 E-mail: myhsiao@ntu.edu.tw
• Received: January 4, 2025 • Revised: June 14, 2025 • Accepted: July 25, 2025
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
To investigate the temporal and kinematic parameters of hyoid bone excursion (HBE) in head and neck cancer (HNC) patients with and without aspiration.
Methods
Videofluoroscopic swallowing study images from 28 HNC patients were divided into aspiration and non-aspiration groups. The kinematic parameters of HBE, including displacement, instantaneous velocity, and instantaneous acceleration, as well as the timing of reaching maximal values in these parameters, were analyzed.
Results
The timings of reaching maximal horizontal (2.37±1.10 seconds vs. 1.09±1.58 seconds, p=0.010; 0.68±0.28 vs. 0.37±0.26, p=0.010 for percentage of time), vertical (1.83±2.06 seconds vs. 0.86±1.42 seconds, p=0.020) and hypotenuse instantaneous velocities (2.36±1.96 seconds vs. 0.79±1.20 seconds, p=0.006; 0.60±0.33 vs. 0.33±0.24, p=0.028 for percentage of time), as well as the timings of reaching maximal horizontal (2.22±1.50 seconds vs. 0.90±1.26 seconds, p=0.009; 0.60±0.32 vs. 0.37±0.29, p=0.041 for percentage of time), vertical (2.09±1.94 seconds vs. 0.83±1.19 seconds, p=0.003), and hypotenuse instantaneous accelerations (2.49±1.93 seconds vs. 0.81±1.24 seconds, p=0.004; 0.65±0.34 vs. 0.34±0.28, p=0.026 for percentage of time) were significantly prolonged in the aspiration group. After signal smoothing, the aspiration group exhibited delayed timing in reaching maximal horizontal instantaneous velocity (2.07±1.09 seconds vs. 0.74±1.10 seconds, p=0.004; 0.58±0.29 vs. 0.32±0.24, p=0.017 for percentage of time), as well as maximal horizontal (2.18±1.16 seconds vs. 1.12±1.46 seconds, p=0.008) and hypotenuse accelerations (2.21±2.50 seconds vs. 0.81±1.21 seconds, p=0.011). There were no significant between-group differences in other kinematic parameters, except for horizontal displacement (7.66±6.26 mm vs. 12.14±5.82 mm, p=0.042).
Conclusion
The timings of reaching maximal instantaneous velocities and accelerations of HBE, rather than the maximum values of these kinematic parameters, may be critical parameters related to aspiration in HNC patients.
Dysphagia is an important clinical issue in patients with head and neck cancer (HNC). Long-term dysphagia affects 46% of patients with HNC, despite improvements in treatments [1]. Dysphagia may cause aspiration pneumonia and malnutrition [2], which have impacts on patient’s survival [3] and life quality [4]. The upward and forward movements of the hyoid bone excursion (HBE) are among the most critical events in the pharyngeal phase of swallowing, related to airway protection and upper esophageal sphincter opening. This process facilitates the entry of the bolus into esophagus and helps avoid pyriform sinus stasis [5-7]. However, in most of previous HBE analyses in videofluoroscopic swallowing study (VFSS), only the maximal displacement and mean velocity of the upward and forward movement of HBE have been investigated, presumed to be possible determinants to predict aspiration. Nevertheless, previous studies have shown inconsistent results in populations with different diagnoses [7-15]. Also, the temporal parameters related to hyoid burst were seldom discussed, which include peak velocity, peak acceleration and timing of reaching these values. The hyoid burst, indicating the rapid abrupt movement of the hyoid in an anterosuperior trajectory, plays an important role for the laryngeal vestibule closure (LVC) [16,17]. A previous study had demonstrated that the timing of reaching peak velocity is closely linked to the timing of achieving LVC in normal population [18]. Recent studies of HBE tracking have also shown that stroke patients with aspiration were associated with a lower maximal horizontal instantaneous velocity and acceleration and prolonged time to these values compared to the non-aspirators [19,20]. Accordingly, these temporal kinematic parameters may be highly associated risk of aspiration.
In a previous kinematic study of HNC patients with dysphagia, it was found that the forward displacement and velocity of HBE in the nasopharyngeal carcinoma (NPC) patients were less than those of healthy subjects. Additionally, the NPC subjects with aspiration experienced less forward displacement of the hyoid bone than those without aspiration [9]. Another study demonstrated a significant decrease not only in the maximal distance but also in the maximal velocity of HBE in both the anterior and the superior direction in postoperative advanced tongue cancer patients who underwent intra-arterial chemoradiotherapy [21]. One study revealed that HNC patients with reduced superior hyoid displacement subjectively perceived more significant swallowing impairment, according to a study-specific subjective swallowing questionnaire. However, kinematic measurements, including anterior and superior hyoid displacements, and temporal measurements, including the hyoid bone elevation onset and duration, were not correlated with the penetration-aspiration score (PAS) score or residual scores [22]. Collectively, these studies suggest that the temporal kinematic parameters of HBE play an important role in dysphagic NPC patients, but inconsistent results have been observed. Furthermore, limited research has investigated the instantaneous kinematics parameters and the timing of HBE in HNC patients. Therefore, the aim of this study is to investigate the temporal kinematic parameters of HBE, including instantaneous velocity and instantaneous acceleration of HBE, between aspiration and non-aspiration groups of HNC patients.
METHODS
Participants and materials
This study collected VFSS videos from 301 patients who presented with swallowing problems at our outpatient clinic and underwent VFSS between May 2015 and May 2020. Only thin liquid test videos from patients diagnosed with HNC were included. If a patient had two examinations within this period, only the latest exam was considered. All the VFSS videos underwent de-identification before data analysis. The trimmed video segment captures the entire hyoid motion during swallowing. Specifically, the kinematic data were extracted from the frame marking the onset of the hyoid burst, defined as the point at which the hyoid bone begins a rapid, abrupt anterosuperior movement, until the end of the excursion, defined as the return of the hyoid bone to its resting position. The VFSS results were categorized into 2 groups: the aspiration group and the non-aspiration group. The aspiration group was defined as PAS in the range of 6–8, while the non-aspiration group included the VFSS images scored as PAS 1–5 [23]. Videos with poor image quality, defined as those with a blurred or poorly visualized hyoid bone, significant motion artifacts, or the presence of metal implants obscuring the region of interest, were excluded. In addition, videos without an obvious swallowing reflex—defined as the absence of a distinct hyoid burst during the swallowing attempt—were also excluded. This retrospective study was approved by the Institutional Review Board of National Taiwan University Hospital (No. 202006027RINC); the requirement for written informed consent was waived because only de-identified data were analyzed, in accordance with institutional regulations and the Declaration of Helsinki.
Twenty-eight candidates were included in this study, with tumor sites among HNC patients encompassing the oral cavity, oropharynx, nasopharynx, hypopharynx, and larynx. The majority of HNC patients were at stage III–IV. All the HNC patients included had undergone concurrent chemoradiation therapy (CCRT). The mean age and height of the HNC patients in these 2 groups were similar. Demographic data is provided in Table 1.
VFSS study
To conduct standardized VFSS, a remote-controlled fluoroscopy with a frame rate of 30 frames per second was utilized (Ultimax-i DREX-UI80; Canon). Subjects were positioned on a customized chair (VFES chair) with a headrest to controlled the head posture, maintaining a fixed distance to the videotape recorder. The patients maintain an upright posture with their head in a neutral position. Each patient ingested 5 mL of a standardized barium sulfate formula (Baritop LV 300 g in 180 mL water) in each single test, using a spoon in most cases. However, for patients with severely impaired lips or tongue movement, a syringe was used when necessary to ensure safe and effective delivery. The thin, thick and paste barium sulfate consisted of 100 mL of Baritop LV suspension and 1.5, 2, 2.5 g of thickener (neo-high toromeal iii). The viscosities (cP=centipoise=10-4 kg·m-1·s-1=10-4 Pa·s) were 0.89, 1,700, and 10,500 cPs, respectively.
Measurement
This study utilized the AetherAI platform (provided by AetherAI Corp.) to identify points of interest and record each parameter during swallowing [20]. Two physiatrists (SHC and KCW) identified six anatomic landmarks (Fig. 1) on the platform, including the anterior-inferior corner of the mandible, anterior-inferior corner of the hyoid bone, and anterior-inferior corner of C2 to C5 vertebrae. These landmarks of C spine vertebrae were employed to transform the coordinate system. The new y-axis was defined by the anterior-inferior corner of C3 and C5, while the new x-axis was perpendicular to the new y-axis and across the anterior-inferior corner of C4. The anterior-inferior C4 vertebrae was established as the origin of the new coordinate system. The coordinates of the hyoid bone position in each VFSS time frame were obtained, and these data were converted from pixel to millimeters (mm). In this coordinate system, anterior displacements were coded as negative values on the x-axis (Fig. 1). To adopt the more intuitive convention where positive values denote anterior motion, we have inverted the sign of horizontal displacement (Dx), so that positive Dx now corresponds to anterior movement, while the vertical displacement (Dy) remains unchanged, with positive values indicating superior motion.
Data processing and statistical analysis
To improve the reliability of signal features such as peak values and timing, displacement signals in both horizontal and vertical directions underwent signal smoothing using a Hilbert–Huang Transform (HHT) based method [24]. First, each displacement signal was decomposed into intrinsic mode functions (IMFs) via empirical mode decomposition. For each IMF combination, the reconstructed signal was smoothed using a third-order Savitzky–Golay (SG) filter with a window size equivalent to one-fifth of the sampling rate (fs=30 Hz), and aligned to the original starting value [25].
To determine the optimal IMF range for reconstruction, a weighted error metric was calculated for each combination, comprising the absolute difference in peak magnitude and the temporal difference of the peak location between the original and the reconstructed signals. The total error was defined as:
The weights for amplitude and timing were fixed at wval=1 and wtime=10, in order to preserve the temporal accuracy of peak displacement while avoiding the amplification of physiologically irrelevant minor oscillations. The IMF range yielding the lowest total error was selected for final signal reconstruction.
Instantaneous velocities and accelerations were computed using the central difference method. For each time point, velocity was approximated by the difference between adjacent displacement values divided by twice the time interval, and acceleration was computed by the second-order symmetric difference formula. The time interval was defined as the inverse of the sampling rate (1/30 s). These calculations were applied to both the original and the smoothed displacement signals. For each subject, the maximal values of displacement, velocity, and acceleration along each axis (Dx, Dy, D, Vx, Vy, V, Ax, Ay, A), and the corresponding time points, were extracted from both the original and smoothed signals. The timing of peak values was also normalized as a percentage of total HBE duration. The representative images of the maximal values of these temporal and kinematic parameters before and after smoothing were demonstrated in Fig. 2.
The results were presented as means and standard deviations. The Mann–Whitney U-test was utilized to analyze the difference in each kinematic and temporal parameter between the aspiration group and the non-aspiration group. All processing and statistical analyses were conducted using MATLAB R2024b (MathWorks Inc.). Statistical significance was considered when the p-value was <0.05.
RESULTS
The main results of kinematic and temporal analysis of HBE are presented in Table 2 and Table 3, as well as Fig. 3. The Dx of the hyoid bone was significantly lower in the aspiration group (7.66±6.26 mm) compared with the non-aspiration group (12.14±5.82 mm) (p=0.042 by Mann–Whitney U-test). There was no significant between-group difference in vertical and hypotenuse displacement, instantaneous velocity, and instantaneous acceleration, in both original and smoothed signals.
In the aspiration group, significant temporal differences were observed compared to the non-aspiration group across key temporal parameters. The timing of reaching maximal horizontal velocity was notably delayed in the aspiration group (2.37±1.10 seconds vs. 1.09±1.58 seconds, p=0.010; 0.68±0.28 vs. 0.37±0.26, p=0.010 for percentage of time). Similarly, the timing of reaching maximal vertical instantaneous velocity showed a significant extension (1.83±2.06 seconds vs. 0.86±1.42 seconds, p=0.020), as did the timing of reaching maximal hypotenuse instantaneous velocity (2.36±1.96 seconds vs. 0.79±1.20 seconds, p=0.006; 0.60±0.33 vs. 0.33±0.24, p=0.028 for percentage of time). Furthermore, the timing of reaching maximal horizontal instantaneous acceleration was significantly longer in the aspiration group (2.22±1.50 seconds vs. 0.90±1.26 seconds, p=0.009; 0.60±0.32 vs. 0.37±0.29, p=0.041 for percentage of time). Correspondingly, the timing of reaching maximal vertical instantaneous acceleration was also prolonged in the aspiration group (2.09±1.94 seconds vs. 0.83±1.19 seconds, p=0.003), as did the timing of reaching maximal hypotenuse instantaneous acceleration (2.49±1.93 seconds vs. 0.81±1.24 seconds, p=0.004; 0.65±0.34 vs. 0.34±0.28, p=0.026 for percentage of time).
In the smoothed signal analysis, several statistically significant differences still noted between groups in temporal parameters. The aspiration group exhibited delayed timing in reaching maximal horizontal instantaneous velocity (2.07±1.09 seconds vs. 0.74±1.10 seconds, p=0.004; 0.58±0.29 vs. 0.32±0.24, p=0.017 for percentage of time). Further delays were observed in the timing of reaching maximal horizontal instantaneous acceleration (2.18±1.16 seconds vs. 1.12±1.46 seconds, p=0.008) and in reaching the peak of hypotenuse acceleration (2.21±2.50 seconds vs. 0.81±1.21 seconds, p=0.011). These findings collectively highlight the notable temporal alterations in HBE within the aspiration group.
DISCUSSION
This study investigated the temporal and kinematic parameters of the hyoid burst in HNC patients. Compared with the non-aspiration group, the timings of reaching maximal velocity and acceleration in the horizontal, vertical and hypotenuse axes, were significantly prolonged in the aspiration group. After signal smoothing, these temporal differences remained observable, although statistical significance was attenuated. Consistent with previous studies, the aspiration group of HNC patients had a lower maximal Dx of the HBE. However, there were no significant differences in instantaneous velocity and acceleration between the aspiration group and the non-aspiration group of HNC patients.
To the best of our knowledge, this is the first study to explore temporal HBE kinematics parameters including instantaneous velocity and acceleration in the HNC patients. Additionally, our results demonstrate that it is the timing of reaching maximal velocity and acceleration rather than the maximal values that is associated with aspiration in this patient group. These findings may enhance the current VFSS-based assessment by providing more sensitive and objective indicators of dysphagia risk. Furthermore, they suggest a potential direction for rehabilitation strategies focusing on improving the coordination and timing of swallowing-related muscle activity, in addition to muscle strength. The identification of such temporal features may serve as a basis for developing biofeedback-based training programs, aiming to improve the precision and synchrony of swallowing function.
To minimize the influence of measurement noise and reduce trajectory irregularities in hyoid kinematic analysis, we applied signal smoothing to the displacement data. Given that the timing of peak velocity and acceleration is highly susceptible to noise, smoothing may enhance the accuracy of these parameters. In this study, we adopted a combination of HHT and SG filtering for data smoothing while minimizing signal distortion. HHT decomposes non-stationary displacement signals into IMFs, allowing flexible reconstruction. However, smoothing can obscure physiologically relevant peaks. To address this, SG filtering was applied to reduce noise while preserving the amplitude and timing of key features. We further calibrated the IMF range using the peak magnitude and temporal difference of velocity and acceleration, retaining only IMFs that preserved essential kinematic information and avoided noise amplification in derivatives.
In the healthy population, a prior study revealed a statistically significant correlation between the timing of reaching peak hypotenuses velocity of the hyoid bone and the timing of the LVC [18]. Another study conducted in a stroke patient group demonstrated greater disorganization in the temporal events of the pharyngeal phase in the aspiration group [11]. While Kendall et al. [26] found no difference in the time required for reaching maximal elevation between post-radiotherapy HNC patients and normal subjects, the onset of hyoid bone elevation was delayed in HNC patients. Additionally, the time the hyoid remained elevated was significantly longer in the patient group. Starmer et al. [27] further demonstrated that higher radiation exposure of the geniohyoid muscle may be related to changes in hyoid kinematic parameters, including a prolonged duration to maximum hyoid elevation, in the post-radiotherapy oropharyngeal cancer patients. Chang et al. [28] studied the trend of abnormal VFSS findings in post-irradiated NPC patients. The proportion of incomplete hyoid elevation and delayed swallowing trigger time both increased over time.
The timing of reaching the maximal velocity and acceleration may reflect a timely hyoid burst, requiring a sequenced and balanced contraction of the suprahyoid and infrahyoid muscles. Delayed timing of reaching maximal velocity and acceleration may have a physiological meaning of impaired protective mechanisms during swallowing. This could be a consequence of interrupted anatomic structures in dysphagic HNC patients after receiving treatments, including surgery and radiotherapy. In healthy individuals, coordinated sequential firing of the suprahyoid and infrahyoid muscles results in smooth hyoid excursion [29,30], and the timing of reaching maximal horizontal velocity typically precedes maximal Dx, reflecting a forceful and coordinated hyoid burst during this process. However, HNC treatments, such as surgical resection and radiotherapy, often cause fibrosis and adhesions in the suprahyoid and infrahyoid musculature, as well as in adjacent soft tissues. These post-treatment fibrotic changes not only stiffen the surrounding soft tissues, but also compress peripheral nerve tracts, leading to denervation of the swallowing muscles and consequently reduce the capacity for active contraction and passive extension [31]. Such changes may introduce passive recoil forces that counteract active muscle contraction and disrupt the normal firing sequence between these muscle groups. As a result, the loss of synchrony between the suprahyoid and infrahyoid muscles may generate an oscillatory hyoid motion pattern, delaying the occurrence of maximal horizontal velocity.
In terms of maximal instantaneous velocity and acceleration, we hypothesized these parameters might be altered in the aspiration group. However, the result of the presenting study did not align with our original hypothesis. A previous study conducted in stroke patients revealed that maximal instantaneous velocity in the horizontal and vertical axes showed no significant difference between the no penetration/aspiration group, penetration group, and aspiration group. However, when compared with the healthy control group, the study found that both horizontal and vertical instantaneous velocity were slower in the stroke group [8]. Accordingly, these two kinematic parameters may have less predictive value for aspiration or non-aspiration in the HNC patient groups. Nevertheless, further investigation of these kinematic parameters in HNC patients compared with healthy subjects and its physiological meaning may be needed.
Previous studies have demonstrated that horizontal hyoid bone kinematics, including displacement and velocity, may be a pivotal role in dysphagia, and this observation is consistent across various disease diagnoses. A study conducted by Steele et al. [15] revealed that reduced anterior hyoid bone displacement was associated with higher penetration-aspiration and pharyngeal stasis. Additionally, Zhang et al. [12] found that maximal anterior movement of hyoid bone might serve as a predictor of penetration-aspiration in patients with dysphagia. In specific diseases, the stroke patient group exhibited decreased maximal horizontal velocity compared to healthy controls [8]. Lee et al. [32] observed that the maximal Dx and velocity of forward hyoid motions were significantly reduced in patients with poor prognosis, defined as necessitates tube placement or diet modification at 6 months after the onset of stroke. In patients with Parkinson’s disease, maximal Dx and velocity significantly decreased during the initial backward and forward motions compared to elderly controls [33].
Previous studies focusing on hyoid bone kinematics in HNC patients with dysphagia have primarily investigated hyoid bone displacement and overall velocity, emphasizing the importance of the horizontal hyoid bone kinematic in dysphagic HNC patients. Wang et al. [9] found that both overall and forward displacement, as well as the overall movement velocity of the hyoid bone in NPC patients were less than those in healthy controls. Moreover, when comparing the aspiration group with the non-aspiration group of NPC patients, they observed a significantly decrease in overall displacement and movement velocity of the hyoid bone in the aspiration group. Cheng et al. [34] demonstrated a remarkable association between anterior hyoid bone displacement and penetration-aspiration in NPC patients, correlating it with the cross-sectional area of the geniohyoid under ultrasonography. Regarding the vertical movement of the hyoid bone, Ishida et al. [35] concluded that upward displacement of the hyoid bone in swallowing was primarily related to events in the oral cavity, while forward displacement was associated with pharyngeal processes, especially the opening of the upper esophageal sphincter. However, a previous study did not show a significant difference in the vertical hyoid bone displacement between irradiated NPC patients and a control group, as well as between aspiration and non-aspiration group of irradiated NPC patients [9]. In the present study, horizontal hyoid bone displacement significantly decreased in the aspiration group, aligning with findings from previous studies. HNC treatments, including surgery and radiotherapy, may induce fibrosis and adhesion of structures such as suprahyoid muscles, infrahyoid muscles, and nearby soft tissues. These pathological changes may lead to reduced anterior hyoid bone displacement, potentially resulting in inadequate airway protection and an increased risk of aspiration [9,36].
This study has several limitations. First, the retrospective design may introduce selection bias and limit control over confounding variables, such as radiation dose, tumor locations and stages, or surgical methods. The patient cohort consisted of HNC patients with varying tumor locations and clinical stages, these differences in disease characteristics and treatment exposure may result in heterogeneous swallowing pathophysiology. Nevertheless, all patients underwent CCRT, a primary contributor to dysphagia in this patient population. Second, while we postulated that involvement and coordination of suprahyoid and infrahyoid muscles may influence the temporal and kinematic features of HBE, we did not evaluate the individual muscular contributions using electromyographic or sonographic assessment. Further studies should focus on delineating the relationship between suprahyoid and infrahyoid muscle activities and HBE. Third, the relatively small sample size may reduce the statistical power and limit the generalizability of the findings. In terms of the VFSS study itself, our hospital employs a 5 mL single bolus volume for each test, and the commercial barium contrast agent (Varibar® E-Z-EM, Inc) is not available. Therefore, the bolus size and consistency in our hospital did not adhere to the standard MBSImP protocol. Therefore, future large-scale studies using standardized VFSS protocols are warranted.
Conclusion
This work quantitatively measured hyoid kinematic and temporal parameters, demonstrating that the timings of reaching maximal instantaneous velocities and maximal instantaneous accelerations were significantly prolonged in the VFSS-confirmed aspiration group in HNC patients. Consistent with findings from previous studies, the Dx of HBE during swallowing were significantly lower in the aspiration group. This study provides a more detailed discussion about hyoid kinematics in HNC patients with dysphagia and suggested possible aspects of hyoid temporal and kinematic study for further investigation in the future.
CONFLICTS OF INTEREST
Ming-Yen Hsiao is an Editorial Board member of Annals of Rehabilitation Medicine. The author did not engage in any part of the review and decision-making process for this manuscript. Otherwise, no potential conflict of interest relevant to this article was reported.
FUNDING INFORMATION
None.
AUTHOR CONTRIBUTION
Conceptualization: Cheng SH, Wang TG, Hsiao MY. Data curation: Cheng SH, Wei KC, Wang TG. Methodology: Cheng SH, Hsiao MY. Formal analysis: Cheng SH, Wei KC. Project administration: Wang TG, Hsiao MY. Visualization: Cheng SH. Supervision: Wang TG, Hsiao MY. Writing – original draft: Cheng SH. Writing – review and editing: Wang TG, Hsiao MY. Approval of final manuscript: all authors.
ACKNOWLEDGEMENTS
The authors would like to thank AetherAI Co., Ltd., Yih-Hsien Kao, MD, Yu-Chen Wang, MS, Jiun-Jen Yang, MD, and Yi-Ya Pan, BS, for their assistance in data collection and analysis. We also sincerely thank Professor Yu-Hsiang Hsu and Ho-Ting Ku, BS, for their valuable guidance on signal smoothing.
Fig. 1.
The points of interests including (1) the anterior-inferior corner of mandible; (2) the anterior-inferior corner of hyoid bone; and (3) the anterior-inferior corner of C2 to C5 vertebrae, were figured out on the AetherAI platform. The landmarks of C spine vertebrae were used for transformation of the coordinate system. The new y-axis was the anterior-inferior corner of C3 and C5, and the new x-axis was perpendicular to the new y-axis and across the anterior-inferior corner of C4. The anterior-inferior C4 vertebrae was defined as the origin of a new coordination system.
Fig. 2.
Representative images of the maximal values of the temporal and kinematic parameters before and after signal smoothing. (A) Before signal smoothing, measurement noise and trajectory irregularities are evident. (B) After signal smoothing, noise is reduced while the amplitude and timing of key features are preserved. Dx max, maximal horizontal displacement; Dy max, maximal vertical displacement; D max, maximal hypotenuse displacement; Vx max, maximal horizontal velocity; Vy max, maximal vertical velocity; V max, maximal hypotenuse velocity; Ax max, maximal horizontal acceleration, Ay max, maximal vertical acceleration; A max, maximal hypotenuses acceleration.
Fig. 3.
Timing of reaching maximal values of the kinematic parameters before and after signal smoothing. (A) Before signal smoothing, significant temporal differences are observed between the aspiration and non-aspiration groups. (B) After signal smoothing, significant between-group differences remain, including the timing of reaching maximal horizontal instantaneous velocity, maximal horizontal instantaneous acceleration, and hypotenuse instantaneous acceleration. Vx max, maximal horizontal velocity; Vy max, maximal vertical velocity; V max, maximal hypotenuse velocity; Ax max, maximal horizontal acceleration; Ay max, maximal vertical acceleration; A max, maximal hypotenuses acceleration. *p<0.05 by Mann–Whitney U-test.
Table 1.
Demographic data of head and neck cancer patients
Characteristic
Aspiration (n=12)
Non-aspiration (n=16)
p-value
Age (yr)
58.9±9.7
64.2±5.6
0.077
Sex
Male
10
15
Female
2
1
Height (cm)
165.3±6.1
164.5±6.4
0.871
Tumor site
Oral cavity
3
8
Oral pharynx
1
0
Nasopharynx
5
4
Hypopharynx
2
3
Larynx
1
1
Tumor stage
I–II
2
1
III–IV
8
14
Unknown
2
1
Post-operation
Yes
5
7
No
7
9
Post-radiotherapy time (yr)
<1
1
6
≥1
11
10
Values are presented as mean±standard deviation or number only.
p-value by Mann–Whitney U-test.
Table 2.
Comparison of temporal and kinematic variables between the aspiration and non-aspiration groups using thin liquid barium
Dx max, maximal horizontal displacement; Dy max, maximal vertical displacement; D max, maximal hypotenuse displacement; Vx max, maximal horizontal velocity; Vy max, maximal vertical velocity; V max, maximal hypotenuse velocity; Ax max, maximal horizontal acceleration; Ay max, maximal vertical acceleration; A max, maximal hypotenuse acceleration.
*p<0.05 by Mann–Whitney U-test.
REFERENCES
1. Hutcheson KA, Nurgalieva Z, Zhao H, Gunn GB, Giordano SH, Bhayani MK, et al. Two-year prevalence of dysphagia and related outcomes in head and neck cancer survivors: an updated SEER-Medicare analysis. Head Neck 2019;41:479-87.
3. Feng MC, Lin YC, Chang YH, Chen CH, Chiang HC, Huang LC, et al. The mortality and the risk of aspiration pneumonia related with dysphagia in stroke patients. J Stroke Cerebrovasc Dis 2019;28:1381-7.
4. Jones E, Speyer R, Kertscher B, Denman D, Swan K, Cordier R. Health-related quality of life and oropharyngeal dysphagia: a systematic review. Dysphagia 2018;33:141-72.
5. Sasegbon A, Hamdy S. The anatomy and physiology of normal and abnormal swallowing in oropharyngeal dysphagia. Neurogastroenterol Motil 2017;29:e13100.
7. Stokely SL, Peladeau-Pigeon M, Leigh C, Molfenter SM, Steele CM. The relationship between pharyngeal constriction and post-swallow residue. Dysphagia 2015;30:349-56.
8. Seo HG, Oh BM, Han TR. Swallowing kinematics and factors associated with laryngeal penetration and aspiration in stroke survivors with dysphagia. Dysphagia 2016;31:160-8.
9. Wang TG, Chang YC, Chen WS, Lin PH, Hsiao TY. Reduction in hyoid bone forward movement in irradiated nasopharyngeal carcinoma patients with dysphagia. Arch Phys Med Rehabil 2010;91:926-31.
11. Saconato M, Leite FC, Lederman HM, Chiari BM, Gonçalves MIR. Temporal and sequential analysis of the pharyngeal phase of swallowing in poststroke patients. Dysphagia 2020;35:598-615.
12. Zhang Z, Perera S, Donohue C, Kurosu A, Mahoney AS, Coyle JL, et al. The prediction of risk of penetration-aspiration via hyoid bone displacement features. Dysphagia 2020;35:66-72.
14. Leonard R, Kendall KA, McKenzie S. Structural displacements affecting pharyngeal constriction in nondysphagic elderly and nonelderly adults. Dysphagia 2004;19:133-41.
15. Steele CM, Bailey GL, Chau T, Molfenter SM, Oshalla M, Waito AA, et al. The relationship between hyoid and laryngeal displacement and swallowing impairment. Clin Otolaryngol 2011;36:30-6.
16. Smaoui S, Peladeau-Pigeon M, Steele CM. Determining the relationship between hyoid bone kinematics and airway protection in swallowing. J Speech Lang Hear Res 2022;65:419-30.
17. Smaoui S, Peladeau-Pigeon M, Steele CM. Variations in hyoid kinematics across liquid consistencies in healthy swallowing. J Speech Lang Hear Res 2021;64:51-8.
18. Nagy A, Molfenter SM, Péladeau-Pigeon M, Stokely S, Steele CM. The effect of bolus volume on hyoid kinematics in healthy swallowing. Biomed Res Int 2014;2014:738971.
21. Ohashi N, Iwai T, Tohara H, Chiba Y, Oguri S, Koizumi T, et al. Swallowing function in advanced tongue cancer patients before and after bilateral neck dissection following superselective intra-arterial chemoradiotherapy for organ preservation: a case-control study. Oral Radiol 2019;35:230-8.
22. Kraaijenga SA, van der Molen L, Heemsbergen WD, Remmerswaal GB, Hilgers FJ, van den Brekel MW. Hyoid bone displacement as parameter for swallowing impairment in patients treated for advanced head and neck cancer. Eur Arch Otorhinolaryngol 2017;274:597-606.
24. Huang NE, Shen Z, Long SR, Wu MC, Shih HH, Zheng Q, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proc R Soc Lond A 1998;454:903-95.
26. Kendall KA, McKenzie SW, Leonard RJ, Jones CU. Timing of swallowing events after single-modality treatment of head and neck carcinomas with radiotherapy. Ann Otol Rhinol Laryngol 2000;109(8 Pt 1):767-75.
27. Starmer HM, Quon H, Kumar R, Alcorn S, Murano E, Jones B, et al. The effect of radiation dose on swallowing: evaluation of aspiration and kinematics. Dysphagia 2015;30:430-7.
28. Chang YC, Chen SY, Lui LT, Wang TG, Wang TC, Hsiao TY, et al. Dysphagia in patients with nasopharyngeal cancer after radiation therapy: a videofluoroscopic swallowing study. Dysphagia 2003;18:135-43.
30. Vaiman M, Eviatar E, Segal S. Surface electromyographic studies of swallowing in normal subjects: a review of 440 adults. Report 1. Quantitative data: timing measures. Otolaryngol Head Neck Surg 2004;131:548-55.
31. McMillan H, Warneke CL, Buoy S, Porsche C, Savage K, Lai SY, et al. Manual therapy for fibrosis-related late effect dysphagia in head and neck cancer survivors: the MANTLE nonrandomized clinical trial. JAMA Otolaryngol Head Neck Surg 2025;151:319-27.
32. Lee WH, Lim MH, Seo HG, Oh BM, Kim S. Hyoid kinematic features for poor swallowing prognosis in patients with post-stroke dysphagia. Sci Rep 2021;11:1471.
33. Lee WH, Lim MH, Nam HS, Kim YJ, Seo HG, Bang MS, et al. Differential kinematic features of the hyoid bone during swallowing in patients with Parkinson's disease. J Electromyogr Kinesiol 2019;47:57-64.
36. Wei KC, Hsiao MY, Wang TG. The kinematic features of hyoid bone movement during swallowing in different disease populations: a narrative review. J Formos Med Assoc 2022;121:1892-9.
Temporal and Kinematic Measurements of Hyoid Bone Excursion in Patients With Head and Neck Cancer
Fig. 1. The points of interests including (1) the anterior-inferior corner of mandible; (2) the anterior-inferior corner of hyoid bone; and (3) the anterior-inferior corner of C2 to C5 vertebrae, were figured out on the AetherAI platform. The landmarks of C spine vertebrae were used for transformation of the coordinate system. The new y-axis was the anterior-inferior corner of C3 and C5, and the new x-axis was perpendicular to the new y-axis and across the anterior-inferior corner of C4. The anterior-inferior C4 vertebrae was defined as the origin of a new coordination system.
Fig. 2. Representative images of the maximal values of the temporal and kinematic parameters before and after signal smoothing. (A) Before signal smoothing, measurement noise and trajectory irregularities are evident. (B) After signal smoothing, noise is reduced while the amplitude and timing of key features are preserved. Dx max, maximal horizontal displacement; Dy max, maximal vertical displacement; D max, maximal hypotenuse displacement; Vx max, maximal horizontal velocity; Vy max, maximal vertical velocity; V max, maximal hypotenuse velocity; Ax max, maximal horizontal acceleration, Ay max, maximal vertical acceleration; A max, maximal hypotenuses acceleration.
Fig. 3. Timing of reaching maximal values of the kinematic parameters before and after signal smoothing. (A) Before signal smoothing, significant temporal differences are observed between the aspiration and non-aspiration groups. (B) After signal smoothing, significant between-group differences remain, including the timing of reaching maximal horizontal instantaneous velocity, maximal horizontal instantaneous acceleration, and hypotenuse instantaneous acceleration. Vx max, maximal horizontal velocity; Vy max, maximal vertical velocity; V max, maximal hypotenuse velocity; Ax max, maximal horizontal acceleration; Ay max, maximal vertical acceleration; A max, maximal hypotenuses acceleration. *p<0.05 by Mann–Whitney U-test.
Graphical abstract
Fig. 1.
Fig. 2.
Fig. 3.
Graphical abstract
Temporal and Kinematic Measurements of Hyoid Bone Excursion in Patients With Head and Neck Cancer
Characteristic
Aspiration (n=12)
Non-aspiration (n=16)
p-value
Age (yr)
58.9±9.7
64.2±5.6
0.077
Sex
Male
10
15
Female
2
1
Height (cm)
165.3±6.1
164.5±6.4
0.871
Tumor site
Oral cavity
3
8
Oral pharynx
1
0
Nasopharynx
5
4
Hypopharynx
2
3
Larynx
1
1
Tumor stage
I–II
2
1
III–IV
8
14
Unknown
2
1
Post-operation
Yes
5
7
No
7
9
Post-radiotherapy time (yr)
<1
1
6
≥1
11
10
Parameter
Aspiration (n=12)
Non-aspiration (n=16)
p-value
Dx max (mm)
7.66±6.26
12.14±5.82
0.042*
Time to Dx max (s)
2.12±1.26
1.19±1.33
0.017*
% Time to Dx max
0.59±0.25
0.48±0.22
0.189
Dy max (mm)
18.97±15.25
15.36±10.25
0.664
Time to Dy max (s)
1.74±1.56
1.02±1.04
0.364
%Time to Dy max
0.46±0.35
0.48±0.22
0.591
D max (mm)
23.54±12.17
19.93±8.16
0.537
Time to D max (s)
1.85±1.47
1.36±1.72
0.192
%Time to D max
0.53±0.35
0.51±0.22
0.864
Vx max (mm/s)
77.45±34.67
100.60±65.00
0.302
Time to Vx max (s)
2.37±1.10
1.09±1.58
0.010*
%Time to Vx max
0.68±0.28
0.37±0.26
0.010*
Vy max (mm/s)
174.89±134.21
122.45±65.77
0.347
Time to Vy max (s)
1.83±2.06
0.86±1.42
0.020*
% Time to Vy max
0.42±0.24
0.34±0.23
0.189
V max (mm/s)
246.30±200.03
158.10±75.11
0.159
Time to V max (s)
2.36±1.96
0.79±1.20
0.006*
% Time to V max
0.60±0.33
0.33±0.24
0.028*
Ax max (mm/s2)
1,399.16±773.01
1,633.26±909.07
0.568
Time to Ax max (s)
2.22±1.50
0.90±1.26
0.009*
%Time to Ax max
0.60±0.32
0.37±0.29
0.041*
Ay max (mm/s2)
3,598.07±3,142.75
2,317.94±1,013.76
0.324
Time to Ay max (s)
2.09±1.94
0.83±1.19
0.003*
%Time to Ay max
0.52±0.25
0.36±0.25
0.157
A max (mm/s2)
4,510.25±4,646.28
2,962.52±1,582.85
0.397
Time to A max (s)
2.49±1.93
0.81±1.24
0.004*
%Time to A max
0.65±0.34
0.34±0.28
0.026*
Parameter
Aspiration (n=12)
Non-aspiration (n=16)
p-value
Dx max (mm)
7.04±5.33
10.79±5.89
0.133
Time to Dx max (s)
2.02±1.37
1.19±1.33
0.054
% Time to Dx max
0.54±0.26
0.48±0.21
0.397
Dy max (mm)
16.56±14.33
14.59±10.16
1.000
Time to Dy max (s)
1.74±1.53
1.25±1.52
0.317
%Time to Dy max
0.46±0.33
0.51±0.24
0.457
D max (mm)
20.17±11.68
18.37±9.29
0.767
Time to D max (s)
1.58±1.55
1.31±1.57
0.591
%Time to D max
0.39±0.28
0.50±0.21
0.133
Vx max (mm/s)
52.88±23.33
72.63±52.11
0.423
Time to Vx max (s)
2.07±1.09
0.74±1.10
0.004*
%Time to Vx max
0.58±0.29
0.32±0.24
0.017*
Vy max (mm/s)
93.11±73.46
82.89±49.48
0.909
Time to Vy max (s)
1.66±2.18
0.79±1.44
0.245
% Time to Vy max
0.38±0.33
0.30±0.24
0.837
V max (mm/s)
119.22±67.29
122.60±64.03
0.873
Time to V max (s)
2.25±2.04
1.26±1.73
0.098
% Time to V max
0.57±0.35
0.45±0.30
0.415
Ax max (mm/s2)
685.53±294.83
775.24±404.56
0.631
Time to Ax max (s)
2.18±1.16
1.12±1.46
0.008*
%Time to Ax max
0.63±0.25
0.45±0.29
0.146
Ay max (mm/s2)
1,366.70±1,436.23
1,165.68±661.25
0.767
Time to Ay max (s)
1.97±2.61
1.29±1.60
0.296
%Time to Ay max
0.43±0.28
0.50±0.29
0.423
A max (mm/s2)
1,644.20±1,341.04
1,585.34±879.55
0.664
Time to A max (s)
2.21±2.50
0.81±1.21
0.011*
%Time to A max
0.54±0.33
0.37±0.27
0.121
Table 1. Demographic data of head and neck cancer patients
Values are presented as mean±standard deviation or number only.
p-value by Mann–Whitney U-test.
Table 2. Comparison of temporal and kinematic variables between the aspiration and non-aspiration groups using thin liquid barium
Values are presented as mean±standard deviation.
Dx max, maximal horizontal displacement; Dy max, maximal vertical displacement; D max, maximal hypotenuse displacement; Vx max, maximal horizontal velocity; Vy max, maximal vertical velocity; V max, maximal hypotenuse velocity; Ax max, maximal horizontal acceleration; Ay max, maximal vertical acceleration; A max, maximal hypotenuse acceleration.
p<0.05 by Mann–Whitney U-test.
Table 3. Comparison of temporal and kinematic variables after signal smoothing between the aspiration and non-aspiration groups using thin liquid barium
Values are presented as mean±standard deviation.
Dx max, maximal horizontal displacement; Dy max, maximal vertical displacement; D max, maximal hypotenuse displacement; Vx max, maximal horizontal velocity; Vy max, maximal vertical velocity; V max, maximal hypotenuse velocity; Ax max, maximal horizontal acceleration; Ay max, maximal vertical acceleration; A max, maximal hypotenuse acceleration.
