Biomedical Engineering Reference
In-Depth Information
To avoid charring close to the electrodes, which would reduce
tissue conductivity, isolate the catheter, and thereby hamper the
ability of achieving a larger necrotic region, internally cooled
applicators or applicators that irrigate their surroundings with
a saline solution are sometimes used.
Excellent reviews on the simulation of RF ablation have been
published [13, 79]. Many simulations and studies have been
performed (e.g., [10, 19, 24, 73, 75, 77, 83, 84, 105, 109, 114, 147,
148, 152, 159]). They consider effects such as tissue damage
and resulting tissue property changes (sometimes assumed to
be continuous, sometimes threshold triggered), temperature-
dependent parameters, cell water evaporation and diffusion,
convective cooling, the impact of blood vessels (by imposing
a convective boundary condition, simulating laminar flow, or
solving the Navier-Stokes equations for the full-flow picture,
e.g., in branching vessels or the heart lumen), state changes (e.g.,
vapor generation and the involved latent heat requirements),
internal cooling, and saline perfusion. Transient effects are only
rarely considered, and many simulations are performed in 2D
considering the rotational symmetry of the ablation catheter.
The lesion size is either estimated using an isotherm (50−55°C)
criterion or an Arrhenius-based tissue damage model. Finite ele-
ment and finite differences methods (sometimes locally refined
with special, conformal voxels) as well as hybrid models (finite
differences for vessels, FEM for tissue) have been applied. The
SAR distribution is usually obtained by solving the quasistatic
equation [159]. Some simulations include a model of an imped-
ance-based source voltage controller.
It was found that vessels can have a strong impact on the tem-
perature distribution, that different models can result in very
different predictions of the vessel influence (see [148]), and that
the influence depends in a complex manner on the distance to
the vessel, the flow rate, the vessel geometry and flow pattern, as
well as the size of the heated region (e.g., [36, 109]).
Many of the previously mentioned publications include
experimental validation. Especially noteworthy is [195], which
experimentally analyzes the processes of cell water evaporation
and diffusion, and presents an empirical relationship between
temperature and water content during RF ablation.
Only a few comprehensive treatment planning software tools
exist [46, 108, 147, 178]. [178] focuses on segmentation, auto-
matic probe placement (which optimizes tumor coverage by
assuming a given, ellipsoidal lesion size and finding the best
position and orientation), and the identification of unaccept-
able access trajectories. No physical simulations are performed,
no risk estimation is provided, and the impact of vessels is only
considered by locally deforming the lesion ellipsoid near major
vessels [179]. [108] and [60] present a very complete treatment
planning tool, but the heating is triggered by a laser source
and not by RF fields. Nevertheless, much of this work can be
applied to RF ablation. In addition, [60] contains important
work on HPC-based real time assessment, MR thermometry,
and control.
7. 1 1 C h a l l e n g e s
Major progress in HTP has been made during the last decade,
resulting in tools such as HyperPlan and HYCAT. Nevertheless,
the following challenges should be addressed to further increase
the reliability of exposure prediction:
• EM tissue parameters are typically based on the values
compiled by [62], which provides frequency-dependent
models of the dielectric parameters for various tissues.
Thermal properties originate from various literature
sources and measurements. These parameters are comple-
mented by models of thermoregulation (both local and
whole body). However, the thermal properties in partic-
ular show a large inter- and intrasubject variability that
is mostly not considered. A future goal should be to find
ways of providing patient-specific parameters (e.g., perfu-
sion based on MRI perfusion maps).
• Temperature monitoring is still needed during treatment
at present. However, it only provides very sparse and
often unreliable (e.g., intraluminal probes) information,
increases the risk of complications in the case of invasive
monitoring, and comes at high financial cost, as in the
case of MRI. Due to the great progress in HTP, it appears
achievable that—at least for fixed tissue geometries—
reliable prediction is feasible supported by very sparse
experimental information.
• Quality assurance is not sufficiently addressed in
hyperthermia, in particular with respect to applica-
tor characterization resulting in an unacceptably large
interapplicator variability. The possibility of positioning
the patient with sufficient precision is often not provided.
Procedures and protocols vary widely between clin-
ics. In the future, quality assurance guidelines should
be extended (efforts are ongoing within ESHO) and
technology provided to overcome the aforementioned
shortcomings.
• Another challenge is clinical integration. Current tools
have limited interfaces with radiotherapy planning
(HYCAT supports RTSTRUCT data). They bring their
own graphical interfaces, are badly interfaced to the
applicator (related partly to regulatory issues), and have
no interface to hospital information systems. All of this
has to be improved in order to obtain acceptance in the
standard clinical environment.
• Automatization needs to be improved to reduce human
error and obtain acceptable treatment planning times. For
this, an extended wizard approach together with more
sophisticated segmentation functionality (e.g., automati-
zation based on prior knowledge) might be the solution.
Automatic segmentation is, however, very difficult to
achieve for hyperthermia due to the many different tis-
sues that have to be distinguished.
