Cervical Vertebral Height Approximates Hyoid Displacement in Videofluoroscopic Images of Healthy Adults

Elevating Patient Care With Deep Learning: High-Resolution Cervical Auscultation Signals for Swallowing Kinematic Analysis in Nasogastric Tube Patients

Patients with nasogastric (NG) tubes require careful monitoring due to the potential impact of the tube on their ability to swallow safely. This study aimed to investigate the utility of high-resolution cervical auscultation (HRCA) signals in assessing swallowing functionality of patients using feeding tubes. HRCA, capturing swallowing vibratory and acoustic signals, has been explored as a surrogate for videofluoroscopy image analysis in previous research. In this study, we analyzed HRCA signals recorded from patients with NG tubes to identify swallowing kinematic events within this group of subjects. Machine learning architectures from prior research endeavors, originally designed for participants without NG tubes, were fine-tuned to accomplish three tasks in the target population: estimating the duration of upper esophageal sphincter opening, estimating the duration of laryngeal vestibule closure, and tracking the hyoid bone. The convolutional recurrent neural network proposed for the first task predicted the onset of upper esophageal sphincter opening and closure for 67.61% and 82.95% of patients, respectively, with an error margin of fewer than three frames. The hybrid model employed for the second task successfully predicted the onset of laryngeal vestibule closure and reopening for 79.62% and 75.80% of patients, respectively, with the same error margin. The stacked recurrent neural network identified hyoid bone position in test frames, achieving a 41.27% overlap with ground-truth outputs. By applying established algorithms to an unseen population, we demonstrated the utility of HRCA signals for swallowing assessment in individuals with NG tubes and showcased the generalizability of algorithms developed in our previous studies. Clinical impact: This study highlights the promise of HRCA signals for assessing swallowing in patients with NG tubes, potentially improving diagnosis, management, and care integration in both clinical and home healthcare settings.

Validity and reliability of the Group for Learning Useful and Performant Swallowing (GLUPS) tool

INTRODUCTION

To validate the Group for Learning Useful and Performant Swallowing (GLUPS), a clinical tool dedicated to videofluoroscopy swallowing study (VFSS).

METHODS

Forty-five individuals were recruited from January 2022 to March 2023 from the Department of Otolaryngology Head and Neck Surgery of University Hospital Saint-Pierre (Brussels, Belgium). Subjects underwent VFSS, which was rated with GLUPS tool by two blinded otolaryngologists and one speech-therapist. VFSS were rated twice with GLUPS within a 7-day period to assess test-retest reliability.

RESULTS

Twenty-four patients and twenty-one controls completed the evaluations. The internal consistency (α = 0.745) and the test-retest reliability (rs = 0.941; p = 0.001) were adequate. GLUPS reported a high external validity regarding the significant correlation with the Penetration-Aspiration Scale (rs = 0.551; p = 0.001). Internal validity was adequate, because GLUPS score was significant higher in patients compared to controls (6.21 ± 4.42 versus 2.09 ± 2.00; p = 0.001). Interrater reliability did not report significant differences in the GLUPS sub- and total score among the independent judges. The mean GLUPS score of individuals without any evidence of VFSS abnormalities was 2.09/23 (95% CI  1.23-2.95), which supported that a GLUPS score ≥ 3.0 is suggestive of pathological VFSS.

CONCLUSIONS

GLUPS is a clinical instrument documenting the abnormal findings of oral and pharyngeal phases at the VFSS. GLUPS demonstrated high reliability and excellent criterion-based validity. GLUPS may be used in clinical practice for the swallowing evaluation at the VFSS.

Radiographic Magnification on Videofluoroscopy: An Important Variable to Consider for Scaled Analyses of Swallowing

PURPOSE

Traditionally, kinematic measures on videofluoroscopy require the use of an external scalar (such as a penny) to transform pixels to absolute distances. Videofluoroscopy is subject to image magnification based on the distance of the feature of interest to the X-ray source. However, the impact of the position/location of the external scalar on swallowing measures is unknown. Our goal was to systematically investigate the accuracy of various common external scalar locations in lateral and anterior-posterior (A-P) view.

METHOD

U.S. pennies were taped to a styrofoam head in three positions (on the left and right lateral neck and in midline submentally). Locations were measured to ensure equal left and right, as well as midline, placement. A metal screwdriver (6 mm in diameter) was inserted into the premanufactured hole that is centrally located at the bottom of the styrofoam head. The head was centered on a medical tray and placed in the middle of a Siemens Alpha C-arm Fluoroscope field. ImageJ was used to measure penny length in pixels (three locations) in both lateral and A-P views. Penny length was known (19.05 mm), and, therefore, used to derive screwdriver size (for each location) for comparison to the actual screwdriver size.

RESULTS

All scalars overestimated the screwdriver size ranging from 6.55 to 7.87 mm, representing a 9%-31% inflation. Scalars closer to the X-ray source had the largest magnification.

CONCLUSIONS

Our results confirm that image magnification of external scalars is a significant source of variability that is currently unaccounted for in the swallowing literature. Recommendations for future research design/measurement methods are provided.

Non-Invasive Sensor-Based Estimation of Anterior-Posterior Upper Esophageal Sphincter Opening Maximal Distension

OBJECTIVE

Dysphagia management relies on the evaluation of the temporospatial kinematic events of swallowing performed in videofluoroscopy (VF) by trained clinicians. The upper esophageal sphincter (UES) opening distension represents one of the important kinematic events that contribute to healthy swallowing. Insufficient distension of UES opening can lead to an accumulation of pharyngeal residue and subsequent aspiration which in turn can lead to adverse outcomes such as pneumonia. VF is usually used for the temporal and spatial evaluation of the UES opening; however, VF is not available in all clinical settings and may be inappropriate or undesirable for some patients. High resolution cervical auscultation (HRCA) is a noninvasive technology that uses neck-attached sensors and machine learning to characterize swallowing physiology by analyzing the swallow-induced vibrations/sounds in the anterior neck region. We investigated the ability of HRCA to noninvasively estimate the maximal distension of anterior-posterior (A-P) UES opening as accurately as the measurements performed by human judges from VF images.

METHODS AND PROCEDURES

Trained judges performed the kinematic measurement of UES opening duration and A-P UES opening maximal distension on 434 swallows collected from 133 patients. We used a hybrid convolutional recurrent neural network supported by attention mechanisms which takes HRCA raw signals as input and estimates the value of the A-P UES opening maximal distension as output.

RESULTS

The proposed network estimated the A-P UES opening maximal distension with an absolute percentage error of 30% or less for more than 64.14% of the swallows in the dataset.

CONCLUSION

This study provides substantial evidence for the feasibility of using HRCA to estimate one of the key spatial kinematic measurements used for dysphagia characterization and management. Clinical and Translational Impact Statement: The findings in this study have a direct impact on dysphagia diagnosis and management through providing a non-invasive and cheap way to estimate one of the most important swallowing kinematics, the UES opening distension, that contributes to safe swallowing. This study, along with other studies that utilize HRCA for swallowing kinematic analysis, paves the way for developing a widely available and easy-to-use tool for dysphagia diagnosis and management.

Autonomous Swallow Segment Extraction Using Deep Learning in Neck-Sensor Vibratory Signals From Patients With Dysphagia

Dysphagia occurs secondary to a variety of underlying etiologies and can contribute to increased risk of adverse events such as aspiration pneumonia and premature mortality. Dysphagia is primarily diagnosed and characterized by instrumental swallowing exams such as videofluoroscopic swallowing studies. videofluoroscopic swallowing studies involve the inspection of a series of radiographic images for signs of swallowing dysfunction. Though effective, videofluoroscopic swallowing studies are only available in certain clinical settings and are not always desirable or feasible for certain patients. Because of the limitations of current instrumental swallow exams, research studies have explored the use of acceleration signals collected from neck sensors and demonstrated their potential in providing comparable radiation-free diagnostic value as videofluoroscopic swallowing studies. In this study, we used a hybrid deep convolutional recurrent neural network that can perform multi-level feature extraction (localized and across time) to annotate swallow segments automatically via multi-channel swallowing acceleration signals. In total, we used signals and videofluoroscopic swallowing study images of 3144 swallows from 248 patients with suspected dysphagia. Compared to other deep network variants, our network was superior at detecting swallow segments with an average area under the receiver operating characteristic curve value of 0.82 (95% confidence interval: 0.807-0.841), and was in agreement with up to 90% of the gold standard-labeled segments.