2021 Fiscal Year Final Research Report

Computer Assisted Diagnosis for Predicting Lung Cancer by Using Large Scale Low-Dose CT Database

Research Project

PDF

Project/Area Number	19K08155
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	Kumamoto University
Principal Investigator	Shiraishi Junji 熊本大学, 大学院生命科学研究部(保), 教授 (30551311)
Co-Investigator(Kenkyū-buntansha)	小林聡金沢大学, 保健学系, 教授 (30313638) 田中利恵金沢大学, 保健学系, 准教授 (40361985) 南哲弥金沢医科大学, 医学部, 教授 (60436813)
Project Period (FY)	2019-04-01 – 2022-03-31
Keywords	未病 / 人工知能 / 機械学習 / 肺がん検診 / CT / コンピュータ診断支援
Outline of Final Research Achievements	In this study, which aims to detect lung cancer in "Mibyo" stage, among the subjects who are continuously undergoing lung cancer CT screening, abnormal findings were discovered in the middle, although it was initially judged to be normal. We constructed the world's first image database of "Mibyo" cases, assuming that the subject's stage in the last year, before the discovery of abnormal findings, was normal. In the initial stage of this study, we constructed an image database of the entire lung cancer CT examination, which was conducted over 11 years with about 23,000 subjects and about 45,000 examinations, and was the first in the world to do so. An image database of cases in a diseased state was extracted. After that, on the premise of detecting lesions at "Mibyo" stage, we tried to distinguish between smokers and non-smokers, and young people and elderly people by deep learning.
Free Research Field	コンピュータ診断支援
Academic Significance and Societal Importance of the Research Achievements	「未病」とは，検査をしても結果に異状がない，もしくは，検査結果に異状があるけれども自覚症状が何もない状態を指し，未病の段階で，適切な治療を行うことで深刻な事態に進行することを防ぐことが可能になる．しかしながら，未病の段階で何らかの異常を見つけて病気を予防するための診断法に関しては未だ不明な点も多く，その解明には新しい視点からのアプローチが必要である．特に呼吸器領域においては，慢性閉塞性肺疾患や，超早期の肺がん，間質性肺疾患を対象とした未病の診断が考えられるが，専門医や放射線科医の読影に強く依存し，研究代表者らの知る限りでは，呼吸器領域における未病の検出の具現化を試みた研究はこれまでにない．