Development of an automatic wound evaluation system by using AI

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K12732
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 90150:Medical assistive technology-related
• Research Institution: Kyorin University
• Principal Investigator: Ohura Norihiko 杏林大学, 医学部, 教授 (40322424)
• Co-Investigator(Kenkyū-buntansha): 飯坂 真司 淑徳大学, 看護栄養学部, 准教授 (40709630)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: 人工知能 / CNN / 画像解析 / 創傷評価 / AI

Outline of Final Research Achievements

The aim of this study was to automate objective segmentation of pathologically tissue in wounds by using convolutional neural networks. By using 400 images of pressure ulcer, two plastic surgeons divided four segments that consisted healthy skin, ulcer, necrosis, and granulation segments one at a time. The CNN was trained by this supervised data. Finally, we evaluated the accuracy of image segmentation by using CNN. In testing, we achieved an area-under-the-curve; AUC of 0.9942, specificity was 0.9931, sensitivity was 0.9783. Both sensitivity and specificity were better than generally published models.

Free Research Field

創傷治癒、難治性創傷、人工知能、

Academic Significance and Societal Importance of the Research Achievements

CNNは、さまざまな画像の分離に応用されているが、創傷領域での報告は少ない。この画像解析システムは、創傷の組織学的分離を高精度で行うことが可能であった。2022年6月の骨太の方針に、医療現場でのDxを加速する基盤となる全国医療情報プラットフォームの創設が盛り込まれることが決定した。今後このようなAIによる画像解析もDxのひとつとなりうる。