2015 Fiscal Year Final Research Report
Computer-aided Detection Scheme for Coronary Artery Stenoses on Whole Heart Coronary MR Angiography
Project/Area Number |
25350565
|
Research Category |
Grant-in-Aid for Scientific Research (C)
|
Allocation Type | Multi-year Fund |
Section | 一般 |
Research Field |
Medical systems
|
Research Institution | Ritsumeikan University (2015) Mie University (2013-2014) |
Principal Investigator |
|
Co-Investigator(Kenkyū-buntansha) |
ISHIDA MASAKI 三重大学, 医学部附属病院, 助教 (10456741)
SAKUMA HAJIME 三重大学, 医学系研究科, 教授 (60205797)
|
Project Period (FY) |
2013-04-01 – 2016-03-31
|
Keywords | 冠動脈MRA / 冠動脈狭窄 / 検出支援システム / 超解像技術 |
Outline of Final Research Achievements |
The purpose of this study was to develop a computer-aided detection (CADe) scheme for coronary artery stenoses on whole-heart coronary magnetic resonance angiography (MRA). Based on the curvatures and the lengths from beginning and bifurcations, the segmented arteries were classified according to the American Heart Association (AHA) classifications. The signal intensity at each voxel of centerline of the segmented artery was defined by the maximum signal intensity within a circle with a radius of 3 voxels centered at the voxel in the original MRA image. While tracing the centerline, the voxel with which the signal intensity showed rapid reduction was finally detected as arterial stenosis. Threshold levels for this detection were selected empirically at each AHA classification by taking into account the variations in the signal intensities for normal coronary artery. With our CADe scheme, a sensitivity of artery stenoses was 87.2% with 1.83 false positives (FPs) per patient.
|
Free Research Field |
医用画像解析
|