Biomedical Engineering Reference
In-Depth Information
one is trying to combine several elements into a single number.
When the sub-scores are physical parameters such as measures of
dose or volume, it is very hard to arrive at appropriate importance
factors for each of them. As already pointed out, it is much easier
to do so when biophysical quantities, such as TCP and NTCP, are
involved. Overall morbidity can then be represented by the combined
impact of the NTCPs of all the irradiated organs and tissues. From
the point of view of probability theory, the “un-complication
probabilities (equal to one minus the complication probabilities) are
multiplicative. Therefore, one might hope to calculate the overall
NTCP as in the following formula.
NTCP
overall 1 2
However, this approach is far too simplistic. The NTCP for a given
end point is in general a function of any predisposing conditions.
Age, diabetes, or a history of tobacco and alcohol abuse are well-
known examples of such predisposing conditions. Less well-known,
unfortunately, is the quantitative impact of these conditions on the
various NTCP's. Then, too, any given NTCP is defined in terms of a
specific end-point. The same organ can, and indeed almost certainly
will, respond in more than one way to irradiation (e.g., early and late
= 1 - [ (1-NTCP )
⋅
(1-NTCP )
⋅
… ]
reaction) - and, therefore, have more than one endpoint. These
endpoints will be of varying gravity.
The process just mentioned of multiplying un-complication probabilities
to arrive at an overall un-complication probability,
cations equally
. However, a particular complication in one compartment
(say, the skin) may, and in this example, certainly will, be of quite dif-
weights all compli-
ferent importance than a complication in another compartment (say, the
spinal cord). The uncomplication probability is, therefore,
an entirely
unrealistic measure of morbidity
. To get a more realistic measure,
each
complication needs at the very least to be weighted by an importance
by no means easy to arrive at such weighting factors.
factor
. It is
Then, there remains the problem of combining the tumor sub-score
(say, TCP) with the normal tissue sub-scores (say, NTCP
overall
). One
approach, often cited, is that of maximizing a quantity termed
“uncomplicated control” which is computed as:
TCP
uncomplicated overall
)
The idea behind this equation is that the goal of radiation therapy is to
maximize the probability of local control of the tumor subject to there
= TCP
⋅
(1-NTCP
