Biomedical Engineering Reference
In-Depth Information
The compensation technique described here is relatively crude. Better
strategies and computational tools are needed to account for
inhomogeneities and for the effects of mis-registration due to motion
or changes in anatomy. New strategies are particularly needed for
IMPT since its score functions have to be evaluated by a computer,
and present algorithms do not reproduce the intelligence of
experienced planners who, for example, choose beam directions
which avoid “bad” approaches so far as compensator design is
concerned.
HU to water-equivalent density conversion
To compute and compensate for the effects of inhomogeneities on
proton beams, one
needs a quantitative
“map” of the patient's
tissues. It is more than
a coincidence that the
growth of proton beam
therapy occurred just
as computed tomo-
graphy (CT) became
available, for CT pro-
vides just such a map.
As they are derived
from
X-ray
trans-
Figure 11.8. Transformation between Houns-
field units and water-equivalent density
(relative stopping power) for protons. Repro-
duced with permission from Schaffner and
Pedroni (1998).
mission measurements,
CT scan data are in
units of relative X-ray
absorption coefficients
(Hounsfield Units, abbreviated HU). However, detailed measure-
ments have shown that, in analogy with the transformation needed
for photon therapy (e.g., Figure 3.5 in Chapter 3), quite satisfactory
conversions to proton stopping powers relative to water can be
derived from the CT data. Such a conversion is shown in Figure 11.8.
The spatial resolution needed for such maps is set by the scale of
multiple scattering which is of the order of a few millimeters; happily,
the resolution of CT data is reasonably well matched to this.
In proton beam therapy it is common to refer to the
water-equivalent
density
of a material or tissue. This is the density of a fictitious
compound with the same chemical composition as water, one
