基于MRI图像的眼眶脂肪体积定量测定方法

目的建立一种新型基于MRI图像简单、快速、定量测量眼眶脂肪体积的方法。方法本文采集甲状腺相关性眼病患者眼眶和标准脂肪模具的自旋回波序列T1WI成像的MRI图像,通过计算单层脂肪截面积,利用体积累加的方法进行体积定量测量,并对操作计时,同时检验测量方法的精确度、信度和可重复性,以及测量者间的一致性。结果本文所建立的新方法可准确测量出眼眶脂肪体积(平均差百分比绝对值均〈1%),耗时短,可重复性高(CV均〈1%),可信度优(ICC=0.996,95%CI:0.985~0.999),一致性高(Bland-Altman差值法分析95%一致性界限:-0.54~0.90)。结论这是一种快速简单测量眼眶脂肪体...

Full description

Saved in:
Bibliographic Details
Published in南方医科大学学报 Vol. 37; no. 9; pp. 1248 - 1251
Main Author 蔡秋月 陈智毅 蒋薇 罗耀升 李章芳 胡世弟 沈洁
Format Journal Article
LanguageChinese
Published 南方医科大学第三附属医院内分泌代谢科,广东广州,510630 2017
Subjects
Online AccessGet full text
ISSN1673-4254
DOI10.3969/j.issn.1673-4254.2017.09.18

Cover

More Information
Summary:目的建立一种新型基于MRI图像简单、快速、定量测量眼眶脂肪体积的方法。方法本文采集甲状腺相关性眼病患者眼眶和标准脂肪模具的自旋回波序列T1WI成像的MRI图像,通过计算单层脂肪截面积,利用体积累加的方法进行体积定量测量,并对操作计时,同时检验测量方法的精确度、信度和可重复性,以及测量者间的一致性。结果本文所建立的新方法可准确测量出眼眶脂肪体积(平均差百分比绝对值均〈1%),耗时短,可重复性高(CV均〈1%),可信度优(ICC=0.996,95%CI:0.985~0.999),一致性高(Bland-Altman差值法分析95%一致性界限:-0.54~0.90)。结论这是一种快速简单测量眼眶脂肪体积的新方法,具有耗时短、精确度高、重复性好、一致性高、可信度优、学习成本低等优势,极具临床推广价值。
Bibliography:44-1627/R
orbital fat volumes; thyroid-associated ophthalmopathy; magnetic resonance imaging; quantitative measurement
Objective To establish a new method for rapid and quantitative measurement of orbital fat volume based on magnetic resonance imaging (MRI) data.Methods We collected MRI data from normalized mold and patients with the diagnosis of thyroid-associated ophthalmopathy (TAO).The cross-sectional areas of the orbital fat on each MR image slice were measured to calculate the fat volume on each slice and then the total orbital fat volume.We recorded the time for completing the measurement and assessed the precision,reliability,repeatability and interoperator variations of the results.Results This MRI data-based method allowed precise measurement of the orbital fat volumes with an absolute value of the mean percentage difference 1%.This method was fast and the results showed a good repeatability (with CVs 1%),a high reliability (ICC=0.996,95% CI: 0.985-0.999) and a high interoperator concordance (95% CI o
ISSN:1673-4254
DOI:10.3969/j.issn.1673-4254.2017.09.18