Wiesinger Florian, Sacolick Laura I, Menini Anne, Kaushik Sandeep S, Ahn Sangtae, Veit-Haibach Patrick, Delso Gaspar, Shanbhag Dattesh D
GE Global Research, Munich, Germany.
GE Global Research, Bangalore, India.
Magn Reson Med. 2016 Jan;75(1):107-14. doi: 10.1002/mrm.25545. Epub 2015 Jan 16.
To investigate proton density (PD)-weighted zero TE (ZT) imaging for morphological depiction and segmentation of cranial bone structures.
A rotating ultra-fast imaging sequence (RUFIS) type ZT pulse sequence was developed and optimized for 1) efficient capture of short T2 bone signals and 2) flat PD response for soft-tissues. An inverse logarithmic image scaling (i.e., -log(image)) was used to highlight bone and differentiate it from surrounding soft-tissue and air. Furthermore, a histogram-based bias-correction method was developed for subsequent threshold-based air, soft-tissue, and bone segmentation.
PD-weighted ZT imaging in combination with an inverse logarithmic scaling was found to provide excellent depiction of cranial bone structures. In combination with bias correction, also excellent segmentation results were achieved. A two-dimensional histogram analysis demonstrates a strong, approximately linear correlation between inverse log-scaled ZT and low-dose CT for Hounsfield units (HU) between -300 HU and 1,500 HU (corresponding to soft-tissue and bone).
PD-weighted ZT imaging provides robust and efficient depiction of bone structures in the head, with an excellent contrast between air, soft-tissue, and bone. Besides structural bone imaging, the presented method is expected to be of relevance for attenuation correction in positron emission tomography (PET)/MR and MR-based radiation therapy planning.
研究质子密度(PD)加权零回波时间(ZT)成像用于颅骨结构的形态描绘和分割。
开发并优化了一种旋转超快速成像序列(RUFIS)型ZT脉冲序列,以实现:1)有效捕获短T2骨信号;2)软组织的平坦PD响应。采用反对数图像缩放(即-log(图像))来突出骨骼并将其与周围软组织和空气区分开来。此外,还开发了一种基于直方图的偏置校正方法,用于后续基于阈值的空气、软组织和骨骼分割。
发现PD加权ZT成像结合反对数缩放能够出色地描绘颅骨结构。结合偏置校正,也获得了出色的分割结果。二维直方图分析表明,对于-300 HU至1500 HU(对应于软组织和骨骼)之间的亨氏单位(HU),反对数缩放的ZT与低剂量CT之间存在强的、近似线性的相关性。
PD加权ZT成像能够可靠且高效地描绘头部骨骼结构,在空气、软组织和骨骼之间具有出色的对比度。除了骨骼结构成像外,该方法有望在正电子发射断层扫描(PET)/磁共振成像(MR)中的衰减校正以及基于MR的放射治疗计划中发挥作用。