2023 Fiscal Year Final Research Report

Image analysis of vestibular schwannoma with clustering techniques

Research Project

PDF

Project/Area Number	20K09371
Research Category	Grant-in-Aid for Scientific Research (C)
Allocation Type	Multi-year Fund
Section	一般
Review Section	Basic Section 56010:Neurosurgery-related
Research Institution	Ehime University
Principal Investigator	Kunieda Takeharu 愛媛大学, 医学系研究科, 教授 (60609931)
Co-Investigator(Kenkyū-buntansha)	山田啓之愛媛大学, 医学系研究科, 准教授 (00403808) 大石直也京都大学, 医学研究科, 特定准教授 (40526878) 中村壮志愛媛大学, 医学部附属病院, 助教(病院教員) (50794468) 井上明宏愛媛大学, 医学部附属病院, 講師 (20593403) 末廣諭愛媛大学, 医学系研究科, 助教 (50775012)
Project Period (FY)	2020-04-01 – 2024-03-31
Keywords	MRI画像解析 / 前庭神経鞘腫 / セグメンテーション / 深層学習 / 予後予測
Outline of Final Research Achievements	Since natural history of vestibular schwannomas is not clarified, it is essential to establish quantitative image diagnosis objectively, for purpose of predicting their prognosis. We have developed a new method for automatic tumor segmentation using a group of cases, in whom tumor progression and clinical course has been followed, as training information. Automatic recognition was estimated as unstable, depending on tumor volume and internal characteristics. In addition to the solid tumor group, we successively repeated deep learning to improve accuracy by adding tumor groups with large tumor volumes and/or with cystic components. Automatic analysis of cases not used for learning process resulted in relatively high accuracy of tumor recognition. As a result, the developed automatic tumor recognition algorithm made it possible to evaluate tumor growth without arbitrariness, automated appropriate image analysis in the outpatient setting, and used the knowledge to predict prognosis.
Free Research Field	脳神経外科学
Academic Significance and Societal Importance of the Research Achievements	前庭神経鞘腫に関する観察研究の多くは画像上での腫瘍最大径を基準とし、腫瘍増大の定義も一定ではなく、自然歴は明らかになっていない。恣意性がなく、自動化が可能な腫瘍認識法を開発した本研究の成果は、客観的な腫瘍体積増大指標だけでなく、多彩な経過をたどる腫瘍の正確な病態と自然歴といった学術的知見につながる。さらには、経過中に治療介入を行った症例についての解析から、治療介入の正当性だけでなく、脳神経系の機能的変容面からの腫瘍進展・進行に関する新たなバイオマーカーの探索が可能となる。最終的には、臨床面で不必要な検査を減らし、最適な治療介入時期を予期することで、医療の効率を最大限にできる先制医療を実現する。