Development of early diagnosis of severe sepsis using pattern recognition analysis of NMR data

2014 Fiscal Year Final Research Report

• Project/Area Number: 26670787
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Emergency medicine
• Research Institution: Kyoto University
• Principal Investigator: SUZUKI Takao 京都大学, 医学(系)研究科(研究院), 講師 (40328810)
• Co-Investigator(Kenkyū-buntansha): KOIKE Kaoru 京都大学, 大学院医学研究科, 教授 (10267164) ; HIRAKAWA Keiko 日本医科大学, 医学部, 助教 (30165162) ; SATO Norio 京都大学, 大学院医学研究科, 講師 (30409205) ; MORIYAMA Tsuyoshi 東京工芸大学, 工学部, 准教授 (80449032)
• Project Period (FY): 2014-04-01 – 2015-03-31
• Keywords: 敗血症診断 / 核磁気共鳴 / パターン認識

Outline of Final Research Achievements

We included sepsis patients aged 20 and over, and diagnosed in emergency department. Their serum samples were measured with Nuclear Magnetic Resonance (NMR). The NMR data were converted to the numerical data by Fourier transform, phase correction, baseline correction, and so on. The data were visualized by principal component analysis, and then analized with Partial Least Squares Discriminant Analysis (PLS-DA). We could see the tendency of separation of “severe sepsis” and “non-severe sepsis” by clustering in the PLS-DA analysis.

Free Research Field

救急・集中治療