Optimization of a Pixel-to-Pixel Curve-Fitting Method for Poroelastography Imaging

详细信息    查看全文

• 作者：Belfor Antonio Galaz ; ^{belfor.galaz@usach.cl} ; ^{bgalazd@gmail.com} ; Rodrigo Hernan Acevedo
• 关键词：Ultrasound elastography ; Poroelastography ; Viscoelasticity imaging ; Rheologic models ; Creep test
• 刊名：Ultrasound in Medicine & Biology
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：43
• 期：1
• 页码：309-322
• 全文大小：2878 K
• 卷排序：43

文摘

Ultrasound poroelastography is an imaging modality used to characterize the temporal behavior of soft tissue that can be modeled as a solid permeated by interconnected pores filled with liquid (poroelastic medium). It could be useful in the stage classification of lymphedema. Generally, time-constant models are applied to strain images, and precision of the fitting process, computational cost and versatility in response to changes in tissues properties are crucial aspects of clinical applications. In the work described here, we performed creep experiments on poroelastic phantoms and used rheologic models to visualize the changes in viscoelastic response associated with fluid mobility. We used the Levenberg–Marquardt algorithm as a fitting tool and performed parametric studies to improve its performance. On the basis of these studies, we proposed an optimization schema for the pixel-to-pixel curve-fitting process. We determined that the bimodal Kelvin–Voigt model describes efficiently the temporal evolution of the strain images in heterogeneous phantoms.