WEKO3
アイテム
A SEMI-AUTOMATED CLASSIFICATION OF VASCULAR COMPONENTS IN MOUSE SOMATOSENSORY CORTEX FROM 3D MULTI-PHOTON LASER SCANNING MICROSCOPIC IMAGE
https://repo.qst.go.jp/records/71118
https://repo.qst.go.jp/records/71118d21d62ff-62fa-4a3a-b63a-5720cd359d60
| Item type | 会議発表用資料 / Presentation(1) | |||||
|---|---|---|---|---|---|---|
| 公開日 | 2013-05-28 | |||||
| タイトル | ||||||
| タイトル | A SEMI-AUTOMATED CLASSIFICATION OF VASCULAR COMPONENTS IN MOUSE SOMATOSENSORY CORTEX FROM 3D MULTI-PHOTON LASER SCANNING MICROSCOPIC IMAGE | |||||
| 言語 | ||||||
| 言語 | eng | |||||
| 資源タイプ | ||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_c94f | |||||
| 資源タイプ | conference object | |||||
| アクセス権 | ||||||
| アクセス権 | metadata only access | |||||
| アクセス権URI | http://purl.org/coar/access_right/c_14cb | |||||
| 著者 |
Kawaguchi, Hiroshi
× Kawaguchi, Hiroshi× Takuwa, Hiroyuki× Tajima, Yousuke× Taniguchi, Jyunko× Ikoma, Youko× Seki, Chie× Masamoto, Kazuto× Kanno, Iwao× Ito, Hiroshi× 川口 拓之× 田桑 弘之× 田島 洋佑× 谷口 順子× 生駒 洋子× 関 千江× 正本 和人× 菅野 巖× 伊藤 浩 |
|||||
| 抄録 | ||||||
| 内容記述タイプ | Abstract | |||||
| 内容記述 | Objectives: In quantitative analyses of positron emission tomography (PET) data, intravascular radioactivity should be diminished considering first-pass extraction fraction of radiotracers for each vascular component including artery, capillary, and vein. The ratio of each vascular component in cerebral blood volume (CBV) are needed for this correction, however, the ratio was assumed to be that in bat wings previously reported. In this study, we developed a semi-automated classification of vascular components in mouse somatosensory cortex from a 3D multi-photon laser scanning microscopy (MPLSM). Methods: A glass cranial window was constructed on a mouse head. After the injection of rhodamine dextran 70kDa, the 3D vasculature was acquired on somatosensory cortex of the anesthetized mouse by MPLSM. The lateral FOV was 488x488mm2 in pixels 0.477x0.477mm2. The z-stack was acquired with 1mm interval from brain surface to 350mm deep. The vessel was extracted from the 3D stack and then classified to vessel components. The vessel region in raw data was blurred in optical axis direction because of the non-focal excitation of fluorescence. In this study, a circuitous approach was employed to solve the problem. This method assumes that the actual vascular shape is locally cylindrical form and the blurred vascular shape has the same centerline with actual shape if the point-spread function is symmetric. The random noise in raw image was reduced by non-local means filter. The denoised image was roughly separated to vessel and the others by discriminant analysis method. The centerline of vessels is obtained by a skeletonize algorithm [3]. The minimum distance, L, was calculated from a centerline location, r, to pixels that have half pixel intensity of r. The vessel is defined as the pixels locating inside of spheres that centers and radius are r and L, respectively. Capillary diameter was defined as less than 6 mm. The arteries and veins are manually estimated from the remained vessel region with referencing number of blanches from penetrating or arising vessels. Results: The figure shows (a) raw data, (b) de-noised image by NLM filter, (c) automatically extracted vessel tree and (d) classified vessel components. The volumetric ratio of vascular component on mouse cortex was 5.5, 7.8 and 86.7% for artery, capillary and vein, respectively, while those on bat wings reported were 15.1, 0.4 and 84.5% [2]. Discussion: The vascular component ratio of mouse cortex is different from that of bat wing. However, note that ratios from this study are just from an example, thus further studies are necessary to fix them. In conclusion, the present technique can estimate the cerebral blood volume with minimal arbitrariness. It can contribute not only quantitative PET analysis but also fundamental researches of microcirculation. References: [1] Mintun et al., J Nucl Med 25(2):177-187(1984), [2] Wiederman, Circ Res 12:375-378(1963). [3] Reniers, D et al, IEEE TVCG 14(2),355(2008). Acknowledgements: This work was partially supported by Konica Minolta Science and Technology Foundation and JSPS KAKENHI Grant Number 22700441. |
|||||
| 会議概要（会議名, 開催地, 会期, 主催者等） | ||||||
| 内容記述タイプ | Other | |||||
| 内容記述 | Brain & BrainPET 2013 | |||||
| 発表年月日 | ||||||
| 日付 | 2013-05-23 | |||||
| 日付タイプ | Issued | |||||
