This poster is published under an
open license. Please read the
disclaimer for further details.
Keywords:
Liver, Ultrasound, MR, Image manipulation / Reconstruction, Technology assessment, Technical aspects, Image registration
Authors:
A. Y. Kim, M. W. Lee, H. K. Lim, Y.-T. Oh, J.-B. Kim, W.-C. Bang, D. K. Shin, S. J. Choi, D. Koh; Seoul/KR
DOI:
10.1594/ecr2015/C-1402
Aims and objectives
Ultrasonography (US) has been most commonly used imaging modality for guidance of percutaneous interventional procedures such as biopsy or local ablation therapy of focal hepatic lesions because of its well-known advantages including real-time guidance,
no radiation hazard,
easy accessibility,
and low cost.1-5 However,
small focal hepatic lesions are sometimes difficult to accurately localize with US because they often have poor conspicuity on US.
4,6,7 This issue draws more attention since small focal hepatic lesions are more frequently detected with the use of state-of-the art computed tomography (CT) and magnetic resonance (MR) imaging.
It can be more challenging to perform precise US-guided interventional procedures for small focal hepatic lesions found in patients with liver cirrhosis when many cirrhosis-related nodules are present near a target lesion and hence,
mistargeting can occur.6,7
To overcome these limitations of US guidance,
fusion imaging has been introduced and showed promising results.4,5,8-14 Fusion imaging combines real-time US images with pre-procedural CT or MR images and displays simultaneously both the real-time US image and the corresponding slice of CT or MR images as the angle and the position of US transducer change.
Therefore,
it can help operators conduct interventional procedures with high confidence and accuracy for target lesions of poor sonographic conspicuity with the real-time,
multi-modality comparison capability.
4,5,8,10,11
However,
fusion imaging techniques introduced by many vendors need several steps to complete image fusion process in a manual manner and thus the process may be cumbersome and time-consuming,
especially for less-experienced operators.
Although the registration time required for image fusion varies depending on the level of operators’ experience and the US machines,
it was reported to take up to as long as 30 minutes.
9,15 Therefore,
the need of easier and faster image fusion of real-time US and pre-procedural CT or MR images in an automatic fashion has been raised to enhance the performance of fusion imaging technique.
16 Although a few vendors have recently introduced automatic image fusion methods,
there have been few literatures dealing with its performance including registration error and time required for image fusion.
In this study,
we developed new fusion imaging methods of real-time US and pre-procedural MR images,
providing both manual and automatic methods.
Since the newly developed manual and automatic image fusion methods are different from those of other vendors,
their performances are unknown.
The purpose of this prospective study is to compare registration error and time required for image fusion between the new manual and automatic methods.