Biomedical Engineering Reference
In-Depth Information
to be harder on normal tissues, and longer times risk increased
tumor proliferation.
Dose and dose-fractionation
The variables here are the total
dose delivered (i.e., the distribution of absolute dose), the dose per
fraction, the number of fractions per day, and the number of rest
days per week (usually, the two weekend days). There is, in
practice, a strong correlation between these factors and overall
time, since short overall treatment times tend to require the
delivery of high doses per fraction and/or more than one fraction
per day. The total dose that a patient can tolerate when the
treatment is given in a single fraction is about a factor of three less
than when eking it out over many fractions, as in established
experience. Total dose is not a good surrogate for effect under all
conditions! (Generally, when a statement of dose is given, the
fractionation scheme should also be indicated.)
Dose-Volume effects
The response to radiation of a tumor
and/or of normal tissues also depends on the distribution of dose
within them. In established experience, one has tended to specify
the delivery of a uniform dose to the tumor, and has tended to
specify a single dose in prescribing constraints on normal tissues.
Biophysical models should, in principle at least, take at least all these
factors into account, for both tumors and normal tissues, so that the
change in the therapeutic ratio can be assessed for any particular
change in practice.
I have spent a significant portion of my life in thinking about
biophysical models
2
and I want to give you my strictly personal view
regarding the above three factors. Simply stated, I regarded the first
two problems listed above as being too difficult for me to be able to
make any useful contribution to their modeling. I have focused my
2
I got involved in biophysical modeling through two experiences. The first
was a comment by a colleague, S. Graffman, when I was worrying a lot
about the possibility that, in proton therapy, inhomogeneities within the
patient might lead to regions of reduced dose that might compromise tumor
control. “Why don't you just try calculating how big an impact that might
have?” asked my friend, who had done similar calculations himself. And
so I did, and was partially comforted. The second experience was in a
working group organized by the NCI in the 1980s to assess the then-new
field of 3D conformal therapy. Members were agonizing about how
accurate, in dose and space, dose calculational algorithms need to be.
“Why don't we just try calculating the biological consequences of any
errors?” I asked. And so we did. I took on TCP, believing it to be simpler,
and J. Lyman was brave enough to take on the modeling of normal tissues.
