Biomedical Engineering Reference
In-Depth Information
reserve of undamaged FSUs needed to continue to function. More
usually, it is assumed that the two organs are independent, and that
the patient can afford to lose one of them - as is usually the case
when surgery is performed. In that case, the overall probability of a
complication is approximately given by the sum of the two separate
complication probabilities.
However, things are not that simple in practice; the damage to the two
organs may be correlated. That is, a damaged organ can cause
additional damage to its irradiated partner. This was demonstrated in
a study in which bilateral kidney irradiation in the mouse was
compared with bilateral kidney irradiation followed 24 hours later by
unilateral nephrectomy (Liao
et al.
, 1994). All assays of renal
damage were less severe in the unilaterally nephrectomized group. In
particular, renal tubule survival was greater in the irradiated and
nephrectomized mice than in those who only were irradiated. Clearly,
communication between paired organs can take place, and such
communication is not taken into account by present models.
Other issues
Finally, I want to make a few additional points. It
deserves to be re-emphasized that any model for NTCP is restricted to
a specific endpoint. The model parameters would be different for a
Grade 2 complication as compared with a Grade 3 complication
of some organ. There may even be different tissue architectures
associated with different endpoints for the same normal tissue.
For parallel architecture organs, the mean dose is often stated to be
the predictive variable of interest - for example, in the cases of the
lung, liver, and parotid glands. While one must always respect the
data, the limited range of techniques used in the clinical studies may
not allow one to disentangle which is the truly predictive variable.
Often, a number of variables are highly correlated with one another. I
find it hard to imagine how the mean dose can be the fundamental
variable. If we believe in the FSU concept, it would mean that a dose
of, say, 20 Gy would have one third of the inactivation potential for
each FSU as would a dose of 60 Gy. This would require an extremely
shallow, and linear, behavior in the dose-response of individual
FSUs.
And last, but by no means least, while the response of normal tissues
to radiation depends on many factors, one of the most important is the
fractionation scheme - i.e., the dose delivered to each point in the
irradiated volume
per fraction
. Models of normal tissue complication
