Development of a new diagnostic technique for idiopathic interstitial pneumonia based on a synthesis of patient images

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K08060
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 52040:Radiological sciences-related
• Research Institution: Fujita Health University
• Principal Investigator: Teramoto Atsushi 藤田医科大学, 保健学研究科, 教授 (00513780)
• Co-Investigator(Kenkyū-buntansha): 塚本 徹哉 藤田医科大学, 医学部, 教授 (00236861) ; 今泉 和良 藤田医科大学, 医学部, 教授 (50362257) ; 齋藤 邦明 藤田医科大学, 保健学研究科, 教授 (80262765)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: 特発性間質性肺炎 / 画像生成 / 画像分類 / CT画像 / 病理画像

Outline of Final Research Achievements

Idiopathic interstitial pneumonia is designated as a rare disease, and accurate diagnosis is necessary for appropriate treatment. In this study, we developed a diagnostic support method for idiopathic interstitial pneumonia by combining image generation, transformation techniques, and deep learning. We used chest CT images and pathological tissue specimen images to extract affected areas and classify idiopathic interstitial pneumonia and general interstitial pneumonia. Additionally, we developed a method to convert and generate CT images of idiopathic interstitial pneumonia from general interstitial pneumonia images. Due to the limited number of specimens for idiopathic interstitial pneumonia, we employed a generative adversarial network (GAN) to generate similar pneumonia images and used them as training data. As a result, high classification accuracy was achieved even with a small number of cases.

Free Research Field

医用画像情報工学

Academic Significance and Societal Importance of the Research Achievements

収集できるデータが少ない医用画像について、画像生成・変換技術の有用性を示すことができた。また、社会的には、特発性間質性肺炎の患者数が少なく診断が困難な現状に対し、画像解析技術の活用により診断の精度と効率を向上させることが期待される。これにより、早期の診断や適切な治療の提供が可能となり、患者の生活の質の向上や医療負担の軽減に寄与することが期待される。