Quantitative analysis of maxillofacial region using texture analysis

2022 Fiscal Year Final Research Report

• Project/Area Number: 21K17101
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 57060:Surgical dentistry-related
• Research Institution: Nihon University
• Principal Investigator: ITO Kotaro 日本大学, 松戸歯学部, 講師 (60868983)
• Project Period (FY): 2021-04-01 – 2023-03-31
• Keywords: Texture analysis

Outline of Final Research Achievements

In 2021, after segmenting the regions on the image of each disease, texture analyses were performed for medication-related osteonecrosis of the jaw, odontogenic maxillary sinusitis, submandibular sialadenitis, and parotid sialadenitis. In addition, texture analysis was performed on changes in the mandibular condyle bone marrow of diabetic patients as a relationship with systemic diseases, and differences in texture parameters were clarified. In 2022, we established differences in MRI texture parameters for each type of vascular malformation. Furthermore, texture analysis was performed using preoperative CT images of cases in which root resorption after orthodontic treatment, and texture features were established as risk factors for root resorption.
After extracting the texture features of each disease and tissue, a receiver operating characteristic curve was created and analyzed to obtain the cut-off value necessary for clinical differentiation.

Free Research Field

歯科放射線学

Academic Significance and Societal Importance of the Research Achievements

本研究から得られた種々のテクスチャパラメータにより、今まで画像診断医の主観に依存していた顎顔面領域の疾患の画像特徴を定量的に表すことができた。定量的なテクスチャパラメータを使用して画像診断を行うことにより、画像診断医の経験年数や能力に左右されずに正確な画像診断を行うことができると考えられる。特に、顎顔面領域のような、解剖学的に非常に複雑で、専門の画像診断医が不足している領域では、多種多様な疾患の定量的な画像診断が可能となることの臨床的意義は非常に大きい。
また、本研究で得られたテクスチャパラメータは近年活発に研究が行われている機械学習を行う上でも指標となる数値となりえる。