Zixu Yan1, Feng Chen2, Fa Wu1, Dexing Kong3. 1. School of Mathematical Sciences, Zhejiang University, Hangzhou, 310027, China. 2. Department of Radiology, First Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, 310003, China. 3. School of Mathematical Sciences, Zhejiang University, Hangzhou, 310027, China. dkong@zju.edu.cn.
Abstract
PURPOSE: The inferior vena cava (IVC) is one of the vital veins inside the human body. Accurate segmentation of the IVC from contrast-enhanced CT images is of great importance. This extraction not only helps the physician understand its quantitative features such as blood flow and volume, but also it is helpful during the hepatic preoperative planning. However, manual delineation of the IVC is time-consuming and poorly reproducible. METHODS: In this paper, we propose a novel method to segment the IVC with minimal user interaction. The proposed method performs the segmentation block by block between user-specified beginning and end masks. At each stage, the proposed method builds the segmentation model based on information from image regional appearances, image boundaries, and a prior shape. The intensity range and the prior shape for this segmentation model are estimated based on the segmentation result from the last block, or from user- specified beginning mask if at first stage. Then, the proposed method minimizes the energy function and generates the segmentation result for current block using graph cut. Finally, a backward tracking step from the end of the IVC is performed if necessary. RESULTS: We have tested our method on 20 clinical datasets and compared our method to three other vessel extraction approaches. The evaluation was performed using three quantitative metrics: the Dice coefficient (Dice), the mean symmetric distance (MSD), and the Hausdorff distance (MaxD). The proposed method has achieved a Dice of [Formula: see text], an MSD of [Formula: see text] mm, and a MaxD of [Formula: see text] mm, respectively, in our experiments. CONCLUSION: The proposed approach can achieve a sound performance with a relatively low computational cost and a minimal user interaction. The proposed algorithm has high potential to be applied for the clinical applications in the future.
PURPOSE: The inferior vena cava (IVC) is one of the vital veins inside the human body. Accurate segmentation of the IVC from contrast-enhanced CT images is of great importance. This extraction not only helps the physician understand its quantitative features such as blood flow and volume, but also it is helpful during the hepatic preoperative planning. However, manual delineation of the IVC is time-consuming and poorly reproducible. METHODS: In this paper, we propose a novel method to segment the IVC with minimal user interaction. The proposed method performs the segmentation block by block between user-specified beginning and end masks. At each stage, the proposed method builds the segmentation model based on information from image regional appearances, image boundaries, and a prior shape. The intensity range and the prior shape for this segmentation model are estimated based on the segmentation result from the last block, or from user- specified beginning mask if at first stage. Then, the proposed method minimizes the energy function and generates the segmentation result for current block using graph cut. Finally, a backward tracking step from the end of the IVC is performed if necessary. RESULTS: We have tested our method on 20 clinical datasets and compared our method to three other vessel extraction approaches. The evaluation was performed using three quantitative metrics: the Dice coefficient (Dice), the mean symmetric distance (MSD), and the Hausdorff distance (MaxD). The proposed method has achieved a Dice of [Formula: see text], an MSD of [Formula: see text] mm, and a MaxD of [Formula: see text] mm, respectively, in our experiments. CONCLUSION: The proposed approach can achieve a sound performance with a relatively low computational cost and a minimal user interaction. The proposed algorithm has high potential to be applied for the clinical applications in the future.
Authors: Rina D Rudyanto; Sjoerd Kerkstra; Eva M van Rikxoort; Catalin Fetita; Pierre-Yves Brillet; Christophe Lefevre; Wenzhe Xue; Xiangjun Zhu; Jianming Liang; Ilkay Öksüz; Devrim Ünay; Kamuran Kadipaşaoğlu; Raúl San José Estépar; James C Ross; George R Washko; Juan-Carlos Prieto; Marcela Hernández Hoyos; Maciej Orkisz; Hans Meine; Markus Hüllebrand; Christina Stöcker; Fernando Lopez Mir; Valery Naranjo; Eliseo Villanueva; Marius Staring; Changyan Xiao; Berend C Stoel; Anna Fabijanska; Erik Smistad; Anne C Elster; Frank Lindseth; Amir Hossein Foruzan; Ryan Kiros; Karteek Popuri; Dana Cobzas; Daniel Jimenez-Carretero; Andres Santos; Maria J Ledesma-Carbayo; Michael Helmberger; Martin Urschler; Michael Pienn; Dennis G H Bosboom; Arantza Campo; Mathias Prokop; Pim A de Jong; Carlos Ortiz-de-Solorzano; Arrate Muñoz-Barrutia; Bram van Ginneken Journal: Med Image Anal Date: 2014-07-23 Impact factor: 8.545