Biomedical Engineering Reference
In-Depth Information
models, it is common to focus attention on their predictions for what
is termed “partial volume irradiation” - which is defined as the
uniform irradiation of a portion of an organ, with zero dose delivered
to the rest of it.
NTCP depends, of course, on a large number of factors, many of them
poorly known. However, in the context of dose-volume effects, and
limiting ourselves to partial volume irradiation for the moment, the
NTCP is primarily considered to be a function of dose and the volume
of irradiated normal tissue. It is common to use the relative volume,
relative that is to the volume
NTCP
NTCP
of the organ or normal tissue
constant Volume
constant Volume
compartment. Thus, one can
construct a surface in 3D that
represents NTCP as a function
of dose and partial volume.
From the 3D representation
one can draw three 2D graphs,
corresponding to the three
possible orthogonal cuts
through the 3D surface, as
NTCP
NTCP
Dose
Dose
NTCP
NTCP
Dose
Dose
Dose
Dose
constant Dose
constant Dose
Volume
Volume
Volume
Volume
Dose
Dose
constant NTCP
constant NTCP
suggested in Figure 5.6. It is
Volume
Volume
worth bearing this decomposi-
tion in mind, because the
important difference between
between models is often best
expressed in just one of these
three graphs.
Figure 5.6. Schematic demonstration
of how the 3D surface of NTCP(d,v)
can be represented by three 2D
graphs in each of which the third
variable is held constant.
The modeling of NTCP was greatly stimulated by the publication by
Withers
et al.
(1998) in which two types of tissue architecture were
introduced and contrasted, namely
serial
and
parallel
. (A third tissue
type with
graded response
was also defined, but it has not received
the attention that the other two architectures have.) These have
formed the main basis of modeling efforts since that time. Their
implications are based on the hypothesis that normal tissues are
comprised of elemental structures, called
function sub-units
or FSUs.
Each FSU performs some function characteristic of the normal tissue,
and damage of the normal tissue is a consequence of damage to its
FSUs. (The nature of the damage needs to be specified, of course.
Tissues may express more than one endpoint and the mechanisms
need not be the same for all endpoints.) The FSU may be simply a
