How Closely do Machine Ratings of Duration of UES Opening During Videofluoroscopy Approximate Clinician Ratings Using Temporal Kinematic Analyses and the MBSImP?
Cara Donohue, Yassin Khalifa, Subashan Perera and 2 others
PMID 32955619WHAT IT FOUND
High-resolution cervical auscultation signals processed by machine learning detected the timing of upper esophageal sphincter opening within a 3-frame human error tolerance for 82.6% of patient swallows.
The system also classified sphincter opening as normal or impaired with 85.7% accuracy compared to clinical ratings.
Key findings
01The machine learning model detected upper esophageal sphincter opening within a 3-frame tolerance for 82.6% of swallows in the patient dataset.
02The model classified upper esophageal sphincter opening as normal or impaired with 85.7% accuracy against clinical ratings.
03The system generalized to an outside dataset of healthy adults not used during training.
STILL TO COME
How it was doneWhat they foundWhat it means for SLPs
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The study did not establish normative cutoffs for what constitutes impaired sphincter opening duration. The clinical ratings used for comparison were skewed toward minimal or no impairment, with no swallows rated as severely impaired. The algorithm was less accurate at detecting sphincter closure than opening. This is preliminary evidence; the technology is not ready for standalone diagnostic use.
Declared interests
The authors declared no conflicts of interest.
The easy way to misread this
Do not interpret this as evidence that high-resolution cervical auscultation can currently replace videofluoroscopy for diagnosis. The study explicitly states the method is not ready for standalone use and that sphincter opening duration alone cannot diagnose dysphagia.