Biomedical Engineering Reference
In-Depth Information
Caveats concerning models of dose-volume effects of normal tissues
Seminars in Radiation Oncology (2001) provides a review of much of
the data on dose-volume effects in normal tissues, including some
cautionary tales.
As stated above, normal tissues differ from tumors both in that they
are internally highly organized, and that they are usually, by choice,
irradiated inhomogeneously. The heterogeneity of treatment tech-
niques is so great that it is hard to speak of established experience in
connection with dose-volume effects in normal tissues. In what
follows, I review a few selected experiments that illustrate the point
that the models of NTCP that have been used to date cannot be the
full story.
“Serial” architecture
The spinal cord is the normal tissue most often cited as an example of
serial architecture. In this connection, some fascinating experiments
have been performed by van der Kogel, Bijl, and their colleagues on
the rat. In one study (Bijl
et al.
, 2003), they irradiated first one short
length, and then two separated short lengths, of the spinal cord,
aiming their highly collimated beam transverse to the cord, so that its
entire cross section was irradiated and the length of the irradiated
section(s) was well known. They used graded doses to measure the
dose needed to produce a 50% chance of leg paralysis, the ED
50
. The
dose-response was very steep, so the ED
50
could be measured rather
accurately. A sample of their results is shown in Table 5.1.
Table 5.1. Selected results from Bijl
et al
. (2003) comparing the
irradiation of two separated sections with that of a single section of
the rat cervical spinal cord.
length(s) irradiated
(mm)
ED
50
(Gy)
95% confidence
interval (Gy)
Single-section irradiation
4
53.7
49-62
8
24.9
22-29
Two-section irradiation
4 + 4
(8 mm separation)
45.4
40-50
