TY - GEN
T1 - MR-guided liver cancer surgery by nonrigid registration
AU - Tang, Songyuan
AU - Wang, Yongtian
PY - 2010
Y1 - 2010
N2 - MR images acquired from open magnetic resonance system have been applied to guide percutaneous puncture for ablation of liver tumors recently. However, the MR images do not always show the tumors clearly because of the lower magnetic field (0.5T) and various different surgical and pathological conditions. In our study, we use preoperative CT images to assist locate tumors by registration. In our method, livers are firstly extracted from the MR and CT images as region of interest (ROI). Since the CT image is acquired several days before surgery operation, there is enough time for surgeons to segment the liver correctly from the CT image manually. During the surgery, surgeons wish that the segmentation of the liver from MR images should be fast and easily. By using active contour, only few initial points are selected manually to extract the liver from the selected MR image. Then affine registration and BSpline-based free from deformation registration are applied to extracted livers' images. The results are evaluated by visual inspection and the overlap of tumors of livers. It shows that the proposed method works effective for liver tumors treatment.
AB - MR images acquired from open magnetic resonance system have been applied to guide percutaneous puncture for ablation of liver tumors recently. However, the MR images do not always show the tumors clearly because of the lower magnetic field (0.5T) and various different surgical and pathological conditions. In our study, we use preoperative CT images to assist locate tumors by registration. In our method, livers are firstly extracted from the MR and CT images as region of interest (ROI). Since the CT image is acquired several days before surgery operation, there is enough time for surgeons to segment the liver correctly from the CT image manually. During the surgery, surgeons wish that the segmentation of the liver from MR images should be fast and easily. By using active contour, only few initial points are selected manually to extract the liver from the selected MR image. Then affine registration and BSpline-based free from deformation registration are applied to extracted livers' images. The results are evaluated by visual inspection and the overlap of tumors of livers. It shows that the proposed method works effective for liver tumors treatment.
UR - http://www.scopus.com/inward/record.url?scp=77956091229&partnerID=8YFLogxK
U2 - 10.1109/MIACA.2010.5528505
DO - 10.1109/MIACA.2010.5528505
M3 - Conference contribution
AN - SCOPUS:77956091229
SN - 9781424480128
T3 - 2010 International Conference on Medical Image Analysis and Clinical Application, MIACA 2010
SP - 113
EP - 117
BT - 2010 International Conference on Medical Image Analysis and Clinical Application, MIACA 2010
T2 - 2010 International Conference on Medical Image Analysis and Clinical Application, MIACA 2010
Y2 - 11 June 2010 through 13 June 2010
ER -