Deep Learning Analysis to Automatically Detect the Presence of Penetration or Aspiration in Videofluoroscopic Swallowing Study

다국어 초록 (Multilingual Abstract)

Background: Videofluoroscopic swallowing study (VFSS) is currently considered the gold standard to precisely diagnose and quantitatively investigate dysphagia. However, VFSS interpretation is complex and requires consideration of several factors. Therefore, considering the expected impact on dysphagia management, this study aimed to apply deep learning to detect the presence of penetration or aspiration in VFSS of patients with dysphagia automatically.
Methods: The VFSS data of 190 participants with dysphagia were collected. A total of 10 frame images from one swallowing process were selected (five high-peak images and five low-peak images) for the application of deep learning in a VFSS video of a patient with dysphagia. We applied a convolutional neural network (CNN) for deep learning using the Python programming language. For the classification of VFSS findings (normal swallowing, penetration, and aspiration), the classification was determined in both high-peak and lowpeak images. Thereafter, the two classifications determined through high-peak and low-peak images were integrated into a final classification.
Results: The area under the curve (AUC) for the validation dataset of the VFSS image for the CNN model was 0.942 for normal findings, 0.878 for penetration, and 1.000 for aspiration.
The macro average AUC was 0.940 and micro average AUC was 0.961.
Conclusion: This study demonstrated that deep learning algorithms, particularly the CNN, could be applied for detecting the presence of penetration and aspiration in VFSS of patients with dysphagia.

참고문헌 (Reference)

1 Kyeong Eun Uhm ; Sook-Hee Yi ; 장현정 ; 천희정 ; 권정이, "Videofluoroscopic Swallowing Study Findings in Full-Term and Preterm Infants With Dysphagia" 대한재활의학회 37 (37): 175-182, 2013

2 Martin-Harris B, "The videofluorographic swallowing study" 19 (19): 769-785, 2008

3 Steele CM, "Sensory input pathways and mechanisms in swallowing : a review" 25 (25): 323-333, 2010

4 Szegedy C, "Rethinking the inception architecture for computer vision" Institute of Electrical and Electronics Engineers 2818-2826, 2016

5 Pauloski BR, "Relationship between manometric and videofluoroscopic measures of swallow function in healthy adults and patients treated for head and neck cancer with various modalities" 24 (24): 196-203, 2009

6 Finestone HM, "Rehabilitation medicine : 2. Diagnosis of dysphagia and its nutritional management for stroke patients" 169 (169): 1041-1044, 2003

7 Mandrekar JN, "Receiver operating characteristic curve in diagnostic test assessment" 5 (5): 1315-1316, 2010

8 Park D, "Normal contractile algorithm of swallowing related muscles revealed by needle EMG and its comparison to videofluoroscopic swallowing study and high resolution manometry studies : a preliminary study" 36 : 81-89, 2017

9 Howard AG, "Mobilenets: efficient convolutional neural networks for mobile vision applications"

10 Shaker R, "Management of dysphagia in stroke patients" 7 (7): 308-332, 2011

11 Lee JT, "Machine learning analysis to automatically measure response time of pharyngeal swallowing reflex in videofluoroscopic swallowing study" 10 (10): 14735-, 2020

12 Park D, "Findings of abnormal videofluoroscopic swallowing study identified by highresolution manometry parameters" 97 (97): 421-428, 2016

13 Ertekin C, "Electrophysiological evaluation of oropharyngeal dysphagia in Parkinson's disease" 7 (7): 31-56, 2014

14 Tan M, "EfficientNet : rethinking model scaling for convolutional neural networks" 97 : 6105-6114, 2019

15 Chang MC, "Effectiveness of pharmacologic treatment for dysphagia in Parkinson's disease : a narrative review" 42 (42): 513-519, 2021

16 Park S, "Effect of the submandibular push exercise using visual feedback from pressure sensor : an electromyography study" 10 (10): 11772-, 2020

17 Park D, "Effect of different viscosities on pharyngeal pressure during swallowing : a study using high-resolution manometry" 98 (98): 487-494, 2017

18 Jiwoon Yeom ; Young Seop Song ; Won Kyung Lee ; 오병모 ; Tai Ryoon Han ; 서한길, "Diagnosis and Clinical Course of Unexplained Dysphagia" 대한재활의학회 40 (40): 95-101, 2016

19 He K, "Deep residual learning for image recognition" Institute of Electrical and Electronics Engineers 770-778, 2016

20 Yamashita R, "Convolutional neural networks : an overview and application in radiology" 9 (9): 611-629, 2018

21 DeLong ER, "Comparing the areas under two or more correlated receiver operating characteristic curves : a nonparametric approach" 44 (44): 837-845, 1988

22 Yu KJ, "Clinical characteristics of dysphagic stroke patients with salivary aspiration: a STROBEcompliant retrospective study" 98 (98): e14977-, 2019

23 Zhang Z, "Automatic hyoid bone detection in fluoroscopic images using deep learning" 8 (8): 12310-, 2018

24 Lee JT, "Automatic detection of the pharyngeal phase in raw videos for the videofluoroscopic swallowing study using efficient data collection and 3D convolutional networks" 19 (19): 3873-, 2019

25 Ahmed Z, "Artificial intelligence with multi-functional machine learning platform development for better healthcare and precision medicine" 2020 : baaa010-, 2020

26 Lundervold AS, "An overview of deep learning in medical imaging focusing on MRI" 29 (29): 102-127, 2019