Abstract Archives of the RSNA, 2014
Wenli Cai PhD, Presenter: Nothing to Disclose
Difei Lu, Abstract Co-Author: Nothing to Disclose
Yin Wu, Abstract Co-Author: Nothing to Disclose
Gordon J. Harris PhD, Abstract Co-Author: Medical Advisory Board, Fovia, Inc
Segmentation of diseased liver remains a challenging task in clinical applications due to the high inter-patient variability of liver shapes, sizes and pathologies caused by cancers or other liver diseases. The purpose of this study was to develop a semi-automated scheme for segmentation of diseased livers with cancers using as little as five user-identified landmarks.
Based on the 3D iterative mesh transformation constrained by 2D optimal contour searching on transversal image planes, we developed a semi-automated scheme for multi-resolution segmentation of diseased livers with cancers on CT image, called iterative mesh transformation. The initial liver mesh is defined using five liver anatomical landmarks identified by users and a set of points from the chest wall detected automatically. Liver mesh is then deformed in a progressive manner by iterations between 3D mesh transformation based on the deformation transfer model and 2D contour optimization using the dynamic-programming algorithm. Forty (40) IV-contrast enhanced hepatic MDCT cases with biopsy-confirmed liver cancers or metastases were used for evaluation of our proposed semi-automated liver segmentation scheme. The MDCT imaging parameters settings were: 2.5–5 mm collimation, 1.25–2.5 mm reconstruction interval, 175 mA tube current, and 120 kVp tube voltage.
In reference to the liver segmentation by manual-contouring of two radiologists, the volumetric size of these 40 cancerous livers ranged from 1079.2 CC to 4652.3 CC, in which the tumor volume percentages ranged from 1.77% to 53.54%. We quantify the accuracy of the proposed liver segmentation scheme by using five metrics: (1) VOE: volumetric overlap error (%), (2) RVD: relative absolute volume difference (%), (3) ASD: average symmetric surface distance (unit: mm) (4) SSD: root-mean-square of symmetric surface distance (unit: mm), and (5) MSD: maximum symmetric surface distance (unit: mm). The performance of the proposed scheme was VOE=5.88, RVD=2.57%, ASD=0.51 mm, SSD=1.05 mm, and MSD=7.12 mm.
Our semi-automated liver segmentation scheme can achieve accurate and reliable segmentation results with significant reduction of interaction time and efforts when dealing with the diseased liver cases.
Our semi-automated 3D liver segmentation scheme can provide an accurate and efficient liver volumetric measurement for diseased livers.
Cai, W,
Lu, D,
Wu, Y,
Harris, G,
Semi-automated 3D Segmentation of Livers Using User-defined Landmarks in CT Images. Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL.
http://archive.rsna.org/2014/14016620.html