Diagnostic Accuracy in Voice and Swallowing Projects

Full DSW SystemRater Reliability from Laryngeal EndoscopyStrobe

Dr. Bonilha has spent over 7 years evaluating the assessment of vocal fold vibration from stroboscopy and high-speed videoendoscopy as a clinician and researcher. During this time, she has developed skills and a knowledge base of common and novel vocal fold vibratory features in both persons with and without voice disorders. As a result of these experiments, Dr. Bonilha learned that there is critically low rater reliability for the judgment of vocal fold vibratory features.

Dr. Bonilha believes that the poor reliability is reflective of the fact that the scales used to judge and report vocal fold vibratory behavior do not consider the number of increments of the vibratory features that the clinicians’ can reliably differentiate. For example, phase closure has been rated on a 3 level scale (predominately open, equally open and closed, predominately closed) and a 100 level scale (percentage of the cycle that is open). Yet there is no scientific evidence to determine whether the clinician can differentiate 3 levels, 100 levels or somewhere in between. When a clinician is asked to differentiate 3 levels, when they can truly differentiate 5 levels, the clinician goes back and forth as to which category cases on the edge of the levels truly fit into and tends to be unreliable in their judgments of some cases. Conversely, if a clinician is asked to differentiate 100 levels when they can only truly differentiate 5 levels, the clinician is not able to be so precise and will be reliable only to +/- 10 levels.

This theoretical framework suggests that a systematic approach to evaluating and adjusting scales, anchors, and training will improve the reliability of experienced and novice visual-perceptual raters. This work has led to the goal of creating a standardized, reliable, and valid method to interpret laryngeal endoscopy recordings.

This work is currently part of Dr. Bonilha’s KL2 grant. She is working on this project through 3 aims:

  1. Assess the current reliability and validity
  2. Systematically modify the scales, anchors, and training for interpreting stroboscopy using a psychometric approach
  3. Preliminarily evaluate the value of a combined a decision making process to improve rater reliability by using visual-perceptual judgments assisted by measurement.

The outcome of this proposal will be solid preliminary data to support this future line of research to improve rater reliability of laryngeal endoscopy.

MRI and MBSImP to Improve Diagnostic Accuracy and Recovery Prediction of Swallowing Impairment Post-Stroke


Post-Stroke Images

Stroke is the leading cause of disability and the third leading cause of death in the United States (CDC). Approximately 64% of stroke patients have dysphagia after their stroke. Dysphagia is associated with increased health care costs and decreased quality of life. Aspiration pneumonia from dysphagia is a leading cause of hospital mortality and the leading cause of rehospitalization after stroke, making it an important health problem. It is still unclear why some patients recover from dysphagia, while others remain significantly impaired with a higher risk for aspiration.

Prior studies have addressed this topic without the benefit of detailed, quantitative assessment of swallowing impairment and sophisticated MRI analysis tools. Dr. Bonilha’s team, by combining the knowledge and cutting edge methods of experts in swallowing and neuroimaging, is applying recent advancements in these areas to address this question. If doctors and scientists better understood the relationship between stroke location and swallowing impairment, the mechanisms underlying dysphagia and dysphagia recovery could be targeted in medical and behavioral therapies.

This work is currently part of Dr. Bonilha’s SCTR Pilot grant which seeks to answer 3 questions that would expand this knowledge: 1) What is the relationship between the stroke location, stroke lesion volume, and the type and severity of physiologic swallowing impairment? 2) How does swallowing impairment change in the months immediately after stroke? 3) Is swallowing recovery related to the site and volume of the stroke, is it related to the swallowing severity score immediately post-stroke alone, or is swallowing recovery related to a combination of both?

Dr. Bonilha hopes that answers to these questions will help clinicians better predict post-stroke swallowing outcomes, plan appropriate therapies and optimize early swallowing interventional approaches.

Measurement Techniques for Laryngeal Sensation and Mucus Aggregation

Complaints of laryngeal sensation related to mucus and habituation of mucus clearing behaviors are widespread in persons with voice disorders. Despite the fact that irregular laryngeal sensation and mucus aggregation are frequent complaints of patients with voice disorders and frequent topics in our clinics, there is not a reliable and valid way to assess these features.

Based on visual examination and patient information, hydration is monitored and alternative clearing behaviors are promoted. The science that underlies these basic clinical practices is not sufficient to base patient care on.

Dr. Bonilha and other voice researchers have developed and continue to refine a method to judge laryngeal mucus aggregation from laryngeal endoscopy. This method has been applied to understanding the difference in mucus aggregation of persons with and without voice disorders.

As part of her R03 grant, Dr. Bonilha is working to strengthen the scientific underpinnings of clinical practice related to mucus aggregation, mucus clearing behaviors, and laryngeal sensation.

Impact of Technical Factors of Modified Barium Swallow Study (MBSS) on Radiation Exposure and Diagnostic Accuracy

There are several technical components of the MBSS that can influence the diagnostic yield. These components include: tasks elicited, contrast material, standardization within and across patients and clinicians, pulse rate, image exposure, recording systems, file formats and compression, and differences in manufacturer settings.

Dr. Bonilha has recently completed two preliminary studies in this line of research under the mentorship of Dr. Bonnie Martin-Harris as part of her KL2 grant training:

  1. Influence of Clinician Experience, Medical Diagnosis, Swallowing Impairment Severity, and MBSImP Protocol on Radiation Exposure Time During MBSS 
    Guidelines and preventive measures have been established to limit radiation exposure time during modified barium swallow studies (MBSS) but multiple variables may influence exam duration. This study examined the effects of the use of a standardized protocol, Modified Barium Swallow Impairment Profile (MBSImP), on radiation exposure time during MBSS. This study also investigated the impact of swallowing impairment severity, medical diagnosis, and clinician experience.
  2. Preliminary Investigation of the Effect of Pulse Rate on Swallowing Impairment as Judged Using the MBSImP
    Continuous (30 pulse per second) fluoroscopy during modified barium swallow studies (MBSS) is believed to allow greater resolution and accuracy in assessment of swallowing function compared to lower pulse rates. This presumed improvement in diagnostic accuracy exposes patients to increased radiation than lower pulse rates. Thus, it is necessary to verify and quantify the improvement gained from the higher pulse rate to determine the risk/benefit ratio. between the 30pps clinical standard and when pulse rates are reduced to 15pps to minimize radiation exposure.

Vocal Fold Vibratory Features as Seen From High-Speed Videoendoscopy and Stroboscopy

During Dr. Bonilha’s PhD studies, post-doctoral year, and first years as an Assistant Professor at University of South Carolina, the research that she was involved in focused on understanding vocal fold vibratory features as imaged from high-speed videoendoscopy (HSV) and stroboscopy. This work was accomplished under the mentorship of Dimitar Deliyski.

Specifically, during her PhD, Dr. Bonilha and her mentor investigated what was normal from both stroboscopy and high-speed videoendoscopy. In every field, it is necessary to have a thorough understanding of normality prior to creating thresholds for abnormality associated with disorders. In this work, they studied mucosal wave, phase asymmetry, regularity, glottal width, vertical motion, open quotient and glottal configuration. They did this using both standard recordings from HSV and stroboscopy and by processing the recordings into files that highlighted the vibratory feature of interest such as: using kymography to assess symmetry and regularity, highlighting the edges and velocity of the vocal folds to assess mucosal wave, and creating a 3D representation of the vocal fold movement to assess vertical motion.

During Dr. Bonilha’s post-doctoral year and her first years as an Assistant Professor at University of South Carolina, she completed the same investigations that she had conducted in vocally-normal speakers in persons with voice disorders.

Dr. Bonilha is continuing to analyze the large amount of data gathered from those projects to improve the clinical evaluation of persons with voice disorders.


© 2012  Medical University of South Carolina | Disclaimer