Method for detecting aspiration based on image processing-assisted B-mode video ultrasonography

  • Miura Yuka
    Department of Gerontological Nursing/Wound Care Management, Graduate School of Medicine, The University of Tokyo Japan Society for the Promotion of Science
  • Nakagami Gojiro
    Department of Gerontological Nursing/Wound Care Management, Graduate School of Medicine, The University of Tokyo
  • Yabunaka Koichi
    Department of Gerontological Nursing/Wound Care Management, Graduate School of Medicine, The University of Tokyo Department of Ultrasound, Katsuragi Hospital
  • Tohara Haruka
    Gerodontology and Oral Rehabilitation, Department of Gerodontology, Graduate School of Medical and Dental Sciences,Tokyo Medical and Dental University
  • Murayama Ryoko
    Department of Advanced Nursing Technology, Graduate School of Medicine, The University of Tokyo
  • Noguchi Hiroshi
    Department of Life Support Technology (Molten), Graduate School of Medicine, The University of Tokyo
  • Mori Taketoshi
    Department of Life Support Technology (Molten), Graduate School of Medicine, The University of Tokyo
  • Sanada Hiromi
    Department of Gerontological Nursing/Wound Care Management, Graduate School of Medicine, The University of Tokyo

Bibliographic Information

Other Title
  • 画像処理を利用したBモード超音波検査法による誤嚥の検出方法の開発

Search this article

Abstract

  We developed a computer-aided image processing technique to assist in the detection of aspirated boluses by clinicians using B-mode video ultrasonography(BV-US)movies. This study used 42 ultrasonography movies from a previous study that investigated the performance of aspirated bolus detection. This image processing technique aimed to assist clinicians in detecting aspirated boluses and the anterior tracheal walls from BV-US movies in which aspirated bolus appearances were emphasized. We combined a method that focused on the region of interest and the correlation coefficient between current and previous images to emphasize the characteristics of aspirated boluses from brightness-adjusted BV-US images. The sensitivity and specificity of this image processing-assisted method were 0.91 and 0.94, respectively, when we used simultaneously collected videofluoroscopy or videoendoscopy results as references. For four silent aspiration movies for which detection did not occur without assistance of image processing, detection was achieved with image processing assistance. Aspirated boluses detected only with image processing assistance had significantly smaller areas, shorter minor axes, longer major axes, and higher average brightness levels than those detected without image processing. This method will contribute to help clinicians provide modification of food viscosity and posture at an early stage for patients with dysphagia.

Journal

Citations (2)*help

See more

Details 詳細情報について

Report a problem

Back to top