Biomedical Engineering Reference
In-Depth Information
Flat-field transmitter
Sensor coils
(b)
(a)
medSAFE controller
Mid-range transmitter
FIGURE 12.3
Images of two leading electromagnetic tracking devices: (a) NDI Aurora (© 2011 Northern Digital Inc. With permission.); (b)
Ascension 3D Guidance medSAFE
TM
device, shown with a selection of transmitters and sensor coils. (© 2011 Ascension Technology Corp. With
permission.)
Electromagnetic tracking devices have found increasing use
in medical applications in the last decade as tracker technolo-
gies have improved. Electromagnetic tracking sensors that can be
embedded in medical instruments typically consist of small coils
wound around a ferrite core. A field generator creates an electro-
magnetic field that induces a small current in the sensor coil as it
is moved within the field. Position and orientation measurements
of the sensor can then be determined within some working vol-
ume. The primary benefit of electromagnetic tracking is that there
is no line-of-sight requirement as the electromagnetic fields are
not significantly attenuated by the human body. The sensor coils
can also be embedded at the tip of medical instruments such as
catheters and needles to directly track tip placement in the clinical
environment. However, electromagnetic tracking can be sensitive
to metal objects in the working space and this fact must be taken
into account when using electromagnetic tracking devices. Two
leading vendors of electromagnetic tracking systems for the medi-
cal market are Northern Digital Inc. and Ascension Technology
Corp., shown in Figure 12.3.
Irrespective of the tracking system used, its purpose is to
provide a 3D Cartesian representation of the position and orien-
tation of markers attached to surgical tools and the patient anat-
omy. Different tracker manufacturers provide device-specific
application programmer interfaces (API) that can be used to
control the tracker from a host computer and acquire the marker
positions. This position information is typically output as posi-
tion in millimeters and orientation in quaternions or Euler
angles. The nature of the clinical application will determine the
number of tools to be tracked simultaneously and whether opti-
cal or electromagnetic tracking is the best choice.
tools with that of the preoperative imaging modality. In the case
of multimodality datasets this can mean the spatial alignment
of two or more imaging modalities to a common reference coor-
dinate. Registration of the tracked tools with the preoperative
image dataset is typically a manual operation requiring land-
marks defined in the preoperative dataset to be identified using
a tracked tool or tracked stylus. The most common registra-
tion method is a paired point registration that aims to mini-
mize the sum of least square error between the two coordinate
frames (Arun et al. 1987; Horn 1987). Paired point registration
is achieved through the use of multimodality markers affixed
to the patient, shown in Figure 12.4, that serve as the common
reference point between tomographic dataset and tracker space.
Rigid registration schemes have been employed in percutane-
ous procedures, where the procedure is conducted on the same
bed and operating environment as used to acquire the preopera-
tive image. In laparoscopic and open surgical procedures, organ
motion due to respiration and displacement of the abdominal
cavity can be significant. In such instances, nonrigid, deformable
registration schemes may be applied, although at present this is
primarily a research interest (Hawkes et al. 2005; Dandekar et
al. 2007). In percutaneous applications, where accurate account-
ing of respiratory motion using deformable registration could be
beneficial in improving needle placement accuracy, it is not clear
how best to present that respiratory motion information to the
physician during instrument placement.
Recent studies have shown that the fusion of abdominal images
from different modalities can improve diagnosis and monitoring
of disease progression (Kuehl et al. 2008; Giesel et al. 2009). As
positron emission tomography (PET) gains in prominence as a
cancer staging modality, the use of image fusion becomes more
prevalent. Image fusion entails the overlaying of multiple imag-
ing modalities in a spatially relevant form, for example, a func-
tional image such as 18-FDG-PET or dynamic contrast enhanced
(DCE) MRI with a preoperative CT dataset. In the case of RFA of
hepatocellular carcinomas, where postoperative outcome shortly
12.2.4 registration
Registration is defined as the aligning of two disparate coordi-
nate frames. In the context of image-guided interventions this
implies alignment of the coordinate frame of the tracked surgical
