Int. J. Radiation Oncology Biol. Phys., Vol. 28, No. 5, pp.

1135-i 142, 1994

Copyright 0 1994 Elsevier Science Ltd
Printed in the USA. All rights reserved
0360-3016194 $6.00 + MI

0 Special Feature- Physics Original Contribution

BORON NEUTRON CAPTURE THERAPY: A MECHANISM FOR ACHIEVING A

CONCOMITANT TUMOR BOOST IN FAST NEUTRON RADIOTHERAPY

GEORGE E. LARAMORE, PH.D., M.D.,* PETER WOOTTON, B.S. (HoN),*

JOHN C. LIVESEY, PH.D.,* DONALD S. WILBUR, PH.D.,* RUDI RISLER, PH.D.,*
MARK PHILLIPS, PH.D.,* JON JACKY, PH.D.,* THOMAS A. BUCHHOLZ, M.D.,*
THOMAS W. GRIFFIN, M.D.* AND STANLEY BROSSARD, M.S.+

*Department
of RadiationOncology,University
of Washington
MedicalCenterRC-08, Seattle,WA 98 195;
and +Munson Medical Center, Traverse City, MI 49684

Purpose: For many years neutron radiation has been used to treat malignant disease both as fast neutron radiotherapy
and as thermal neutron induced boron neutron capture therapy (BNCT). To date, these two approaches have been
used independently of one another due to the large difference in neutron energies each employs. In this paper we
discuss the potential application of BNCT to enhance the therapeutic effectiveness of a fast neutron radiotherapy
beam.
Methods and Materials: Measurements are presented for the thermal neutron component that is spontaneously
developed as the University of Washington fast neutron radiotherapy beam penetrates a water phantom. The
biological effect of this thermalized component on cells “tagged” with boron-10 (“B) is modeled mathematically
and the expected change in cell survival calculated. The model is then extended to estimate the effect this enhanced
cell killing would have for increased tumor control.
Results: The basic predictions of the model on changes in cell survival are verified with in vitro measurements
-the V-79 cell line. An additional factor of lo-100 in tumor cell killing appears achievable with currently
available l”B carriers using our present neutron beam. A Poisson model is then used to estimate the change in
tumor control this enhanced cell killing would produce in various clinical situations and the effect is sufficiently
large so as to be clinically relevant. It is also demonstrated that the magnitude of the thermalized component can
be increased by a factor of 2-3 with relatively simple changes in the beam generating conditions.
Conclusion: BNCI may provide a means of enhancing the therapeutic effectiveness of fast neutron radiothearapy
in a wide variety of clinical situations and is an area of research that should be aggressively pursued.

Fast neutron radiotherapy, BNCT, Neutron capture, Boron.

INTRODUCTION clinical trials for human malignancies were carried out in

the 1950s using a reactor beam at the Brookhaven Na-
Shortlyafter the neutron was discovered by Chadwick in tional Laboratories to treat malignant gliomas of the brain.
1932, its potential use in the treatment of human malig- In these initial trials boric acid derivatives were used as
nancies was realized. In 1936 Lecher (24) proposed treat- the “B carrier agent (7, 9). Because of the attenuation of
ing tumors with low energy thermal neutrons mediated the thermal neutrons in tissue and problems with necrosis
via a capture cross section reaction with “B. These low of the calvarium, subsequent trials in the 1960s at the
energy thermal neutrons would deposit very little energy Massachusetts Institute of Technology Research Reactor
in passing through most substances but would react used craniotomies both to remove as much tumor as pos-
strongly with selected atomic species such as “B and hence sible and to expose the tumor bed directly to the slow
offer a mechanism for delivering a high energy deposition neutron beam. Unfortunately, these trials showed no
only to cells that were selectively “tagged” with “B. This therapeutic benefit from this form of treatment and
approach was termed boron neutron capture therapy moreover, showed considerable damage to the endothelial
(BNCT) and a few years later Kruger ( 16) and Zahl et al. linings of the blood vessels ( 1,3). This was attributed both
(35) experimentally verified this concept and demon- to the high concentration of l”B in the blood at the time
strated the biological effectiveness of the fission fragments of treatment and to the relatively poorly penetrating char-
produced by slow neutrons interacting with “B. The first acteristics of the collimated reactor beams used for treat-

Reprint requests to: George E. Laramore, Ph.D., M.D. Accepted for publication 22 September 1993.

1135
1136 I. J. Radiation Oncology 0 Biology 0 Physics Volume 28. Number 5, 1994

ment. In 1968 Hatanaka et al. (12. 14) initiated Japanese enhance the killing of tumor cells relative to adjacent nor-
clinical trials of BNCT using Na2BlzH, ,SH (i.e., BSH) as mal tissue cells if one can find a way of selectively deliv-
a carrier agent. Trials have continued to the present time ering ‘“B or another agent with a high thermal neutron
and the data indicates several longterm “cures” of patients cross section to the tumor cell-precisely the same hurdle
with documented glioblastoma multiforme (13). Trials that must be overcome in conventional BNCT using re-
have also been initiated in Japan for malignant melanoma actor beams. There are several potential advantages to the
using a boronated form of phenylalanine as the boron fast neutron augmentation approach: (a) It builds on a
carrier (26). set of well-established, hospital-based, fast neutron radio-
In parallel with this work, other investigators began to therapy facilities that are located in major metropolitan
use high energy neutrons generated by a cyclotron as an- areas and are experienced in conducting clinical trials. (b)
other form of external beam radiotherapy (32). Like the With fast neutron beams it is possible to treat many more
situation with BNCT, the early clinical trials showed con- tumor systems than the relatively superficial tumors to
siderable toxicity and little efficacy and the field languished which low energy reactor beams are restricted. (c) Since
until the 1950s when mammalian cell culture techniques BNCT will be used as a “concomitant boost” for the fast
revealed critical differences between neutron and photon neutron beam, the enhanced tumor cell killing can be
postirradiation cell survival curves. Clinical trials were treated statistically and it may not be necessary to “tag”
resumed at Hammersmith Hospital in London, England almost every tumor cell with “B which is an overwhelm-
in the 1960s and since that time over 15,000 patients have ing problem in conventional BNCT. (d) Normal tissue
been treated with fast neutrons for various malignancies. tolerance doses have already been well-established for fast
This immense data base has enabled us to estimate with neutron radiotherapy which provides a means of esti-
some accuracy the tolerance of most clinically relevant mating the safe dose range for future clinical trials.
normal tissues to fast neutron radiotherapy ( 17, 18). Based In the following sections, we will describe work to date
upon randomized clinical trials and single-institution ex- on characterizing the thermalized component of the fast
periences, fast neutrons appear to offer a therapeutic ad- neutron radiotherapy beam at the University of Wash-
vantage compared to conventional, megavoltage photon ington for a variety of beam generating conditions, discuss
irradiation for the following tumor systems: salivary gland the mathematical modeling used to estimate the biological
tumors (1 l), locally advanced prostate cancer (21, 29) effect of the thermalized component, present some pre-
and sarcomas of bone and soft tissue ( 19). In a randomized liminary biological data which confirms the predictions
trial for patients with inoperable salivary gland tumors, of the model, and then extrapolate this biological effect
initial tumor clearance rates at the primary site were 85% to a projected improvement in tumor control probability.
for neutron-treated patients compared to 33% for photon-
treated patients 0, = 0.0 l), respective local-regional con-
METHODS AND MATERIALS
trol rates at 2 years were 67% vs. 17% (a = 0.005) and 2
year survivals were 62% vs. 25% (p = 0.10). These results Bmm measurement data
were consistent with historical data. Ten-year data con- The problem in using a fast neutron beam to activate
tinue to show a therapeutic advantage to fast neutrons for the BNCT process is that apart from resonance effects,
salivary gland tumors (20). In the case of locally advanced the absorption cross section of “B (and other elements)
prostate cancer, a randomized, clinical trial compared a scales approximately as E-“2 where E is the neutron en-
combination of neutrons and photons (i.e., mixed beam) ergy (4). This makes the process less effective at high neu-
with photons alone. At 10 years, local control rates were tron energies-that is, thecrosssection fora 1 MeV neutron
63% for the mixed beam group compared to 52% for the is about 6000 times smaller than that for a thermal neu-
photon group 0, = 0.05) and respective survival rates were tron. However, a thermalized component is produced
42% vs. 27% (r, = 0.05) (2 1). In the case of sarcomas there spontaneously as the neutron beam penetrates tissue-
have been no randomized clinical trials but a historical the key question is “how large is this thermalized com-
comparison between fast neutrons and conventional ponent?’ To a first approximation we can discuss the
photon irradiation for inoperable tumors shows respective matter in terms of a “fast” component and a “slow” com-
local control rates of 53% vs. 38% for soft tissue sarcomas, ponent. The University of Washington Clinical Fast Neu-
55% vs. 2 1% for osteogenic sarcomas, and 49% vs. 33% tron Radiotherapy System uses a 50.5 MeV p-Be re-
for chondrosarcomas ( 19). For other tumor systems such action to produce the neutron beam used in therapy. The
as squamous cell tumors of the head and neck, esophageal characteristics of the “fast” primary beam have been pre-
carcinomas, and high grade gliomas of the brain, fast neu- viously described (2). The design of the treatment system
tron radiotherapy has exhibited no improvement over readily permits the study of the influence of particle type
conventional photon irradiation. and energy, target thickness, flattening filter conditions,
It now may be appropriate to combine these two tech- and collimator configuration on the thermalized neutron
nologies. As a fast neutron beam enters the body it au- component effective for BNCT. We have measured the
tomatically produces a thermalized component. It should total dose of “fast” neutrons, DN + 7, in a water phantom
be feasible to use this thermalized component to selectively using the European-US protocol (25). The “slow” neutron
 BNCT in fast neutron radiotherapy 0 G. E. LARAMORE et al. 1137

component was measured by activating 23Na in 1 ml al-

iquots of a NaN03 solution (350 g/L). The dose that would
be effective in the BNCT process, Dg, was calculated from
the ratio of the respective cross sections following the pro-
cedure of Waterman et al. (34) and subsequently expressed
in Gy per pg of “B per gram of tissue.
Figure 1 shows the ratio of the “slow” dose to the “fast”
dose along the central axis the indicated beam generating
conditions. Note the scale factor on the vertical axis. The
slow neutron dose is only about 0.1% of the total beam Center Displacement
energy (per pg/g of “B). This ratio peaks at about 6-7 cm
Fig. 2. Beam intensity profiles for the “fast” and “slow” beam
depth when a p+Be reaction is used to generate the neu- components for a 30 cm X 30 cm field generated using a 50.5
tron beam and at about 10-l 1 cm depth when a d+Be MeV p-+Be reaction. The curves were measured at a 5 cm depth
reaction is used. As can be seen from the figure, the overall in a water phantom. The geometry of the flasks used in the
shape of the curve is most strongly determined by the radiobiological experiments is shown schematically in the figure.
particular particle reaction used while the overall mag-
nitude of Dn is dependent on parameters such as field
size and target geometry. We note that Da can be easily imental data sets will be required to model the physical
varied by factors of 2-3 with simple changes in the beam properties of the “fast” and “slow” components.
generating parameters thus indicating that some degree Given that the overall size of the DB component is so
of optimization is possible. It will be important to tailor much smaller than the total neutron dose used in fast
the overall mix of “fast” and “slow” neutrons to a given neutron radiotherapy, the question is whether its biological
clinical situation and hence we expect that “different” effect is sufficient to warrant clinical interest.
beams will be used for superficial vs. deeply seated tumors.
Figures 2 and 3 show beam intensity profiles for the
“fast” and the “slow” components of the beam for 30 cm RESULTS
X 30 cm and 10 cm X 10 cm fields, respectively, using Model calculations of enhancement efect
the standard 50.5 MeV p+Be reaction which produces Model calculations provide a way of estimating the
the fast neutron beam for radiotherapeutic treatments. magnitude of the anticipated, biological effect of BNCT
Note that while the profile of the “slow” neutron dose augmentation of our fast neutron radiotherapy beam. The
mimics that of the “fast” neutron dose reasonably well cell survival probability following a course of fractionated
for the larger field size, it tends to fall off much more neutron radiotherapy can be written as:
rapidly for the smaller field size. Hence, different exper-
”
S = n [e-aDN+T~ IHa+, + At,lnZIrd
1
i=l
X [e- APuBNBDB], (Eq. 1)

where the first term in brackets represents the time-de-

pendent linear-quadratic (LQ) formula for the cell killing
from the primary beam and the second bracketed term

2
0 4 16 20

Fig. 1. Depth dose distribution of the ratio of the thermalized

beam component, Ds, to the primary beam, D,,, along the cen-
tral axis. The particular beam generating conditions and field
sizes used in the measurement are indicated in the figure. “Std”
Center Displacement
refers to the standard target thickness [p (50.5 MeV) + Be (25
MeV)] and “Thk” refers to a thick target in which the charged Fig. 3. Beam intensity profiles for the “fast” and “slow” beam
particle beam is completely stopped. Standard flattening filters components for a 10 cm X 10 cm field generated using a 50.5
used in radiotherapy treatments were used for the standard MeV p-+Be reaction. Unlike Figure 2, the “slow” neutron com-
thickness target. Except as indicated, no filter was used on the ponent is normalized to its maximum value. The curves were
extracted neutron beam from the thick target. measured at a 5 cm depth in a water phantom.
1138 1.J. Radiation Oncology 0 Biology 0 Physics Volume 28. Number 5, 1994

represents the enhancement from the thermalized neutron

component in the presence of “B. In Eq. I, 01 and p are
the usual LQ coefficients, At; is the time interval between
the ith and the i-lth radiation fractions, n is the total
number of fractions, Td is the effective tumor cell doubling
time, A is a parameter of order unity that relates to the
position of the “B atoms in the tumor cells (15) P is the
probability of a neutron capture event fission process re-
sulting in the killing of the cell, @ is the neutron capture
cross section, Ns is the number of l”B atoms on the cell,
I I I I I I I
and DN+? and Ds are, respectively, the “fast” and “slow” -2.0 '
2 3 4 5 6
-1 0 1
neutron components. The enhancement effects are con-
NB (10’ per cell)
tained entirely in the second term since we are dealing
with “B concentrations sufficiently small so as not to de- Fig. 4. Expected enhancement oftumor cell killing as a function
plete the primary neutron beam. To estimate the values of r”B concentration assuming a uniform distribution of “B
of the relevant parameters for the V-79 cell line we use throughout the cell during the entire course of fast neutron ra-
diotherapy. The vertical axis shows the logarithm of the change
the radiobiological measurements and Monte Carlo cal- in cell survival and the horizontal axis shows the r”B concen-
culations of Gabel et al. (8). They first determined the tration in the cells. A standard total dose of 20.4 Gy,, was as-
probability function for a given amount of energy being sumed. The curve for HSEF( 1) is shown as the solid line and
deposited in the cell nucleus from the fission fragments for HSEF(2) is shown as the dotted line.
from a capture event assuming a uniform distribution of
“B and then an al y zed their cell survival data in terms of
this energy deposition function. They then postulated a ment will hence be utilized as a concomitant boost rather
hit size effectiveness function (HSEF) to “map” the energy than being the only mechanism for tumor cell killing.
distribution function onto the experimental cell killing This may simplify somewhat the delivery requirements
probability function. They found that two somewhat dif- in that it may not be necessary to “tag” almost every
ferent HSEFs were consistent with their experimental data. tumor cell with “B.
One function, HSEF(l), was a step function having its
threshold at about 5 keV/p’ and the other, HSEF(%), was Radiobiological measurements
a ramp function with a linear rise between about 2 and To test the predictions of the model calculations we
8 keV/p’. These functions were determined under the performed in vitro measurements of the enhanced cell
assumption that the relative biological effectiveness (RBE) killing for “B added to V-79 cultures. The cells were grown
factors were constant along the fission fragment tracks. in tissue culture flasks in Eagles basal medium supple-
These functions basically show that events that deposit a mented with 10% fetal bovine serum. The cells were then
small amount of energy in the cell nucleus are not effective treated with either cesium mercaptoundecahydro-closo-
in killing the cell while above a certain threshold depo- dodecaborate (99% “B @ 0.8 mM for 42 h) or boric acid
sition energy, an event will result in the cell being killed (99+% “B @ 14 or 28 mM for 1 h). Assuming that these
essentially 100% of the time. Folding the energy deposition treatments would result in a uniform distribution of r”B
function and the HSEF together, we can calculate the throughout the cells, the respective concentrations would
probability, P, in Eq. 1. For HSEF( l), D = 0.846 and for be 100, 250, and 500 pg/g of cellular material. The cells,
HSEF(2), p = 0.782. Assuming a uniform distribution of still within the “B-containing medium, were then irra-
“B throughout the cells, “A” in Eq. 1 is unity and we diated in a water phantom at a depth of 5-6 cm to provide
can evaluate the expected enhanced cell killing as a func- a moderated thermal neutron component as would occur
tion of the number of “B atoms per cell. The resulting naturally in tissue. The flask geometry used in the exper-
curves corresponding to the two different HSEFs are iments is shown schematically in Figure 2. Subsequent
shown as SB(1) and SB(2) in Figure 4. In making the plating of the cells and assaying for the formation of col-
calculation, the total dose of fast neutrons was assumed onies resulted in the survival curves shown in Figure 5.
to be 20.4 Gy,, which is the total dose used in the treat- Unirradiated cells incubated in the same media were used
ment of most tumors based upon prior clinical trials (10). as controls to determine the plating efficiencies which were
The two models for HSEF give essentially equivalent re- used to normalize the respective curves. No significant
sults. While there is no definite consensus on the achiev- toxicity was noted due to the media themselves. These
able tumor concentration of “B using current carrier curves clearly show the enhancement effect of BNCT due
agents, a value on the order of a few times 10’ per cell is to the thermalized component of the fast neutron radio-
a reasonable estimate according to the published literature therapy beam. The effect is somewhat exaggerated due to
(5, 6). Thus, it appears that one can expect an additional the large amount of “B used in the experiments. However,
factor of lo-100 in cell kill if BNCT is used to enhance the maximum radiation dose used was only about half
conventional fast neutron radiotherapy. BNCT enhance- the “standard” neutron radiation dose used in a curative
 BNCT in fast neutron radiotherapy 0 G. E. LARAMORE etal. 1139

0 2 4 6 8 10 12
Do80WY) 0 5 10 15 20 25 30
Fig. 5. In vitro measurements of the BNCT enhanced cell killing ~0s E (Qyv,, I
for V-79 cells using the standard fast neutron radiotherapy beam
Fig. 6. Schematic illustration of dose response curve for a typical
at the University of Washington. The “B carrier compounds
tumor system treated with fast neutron radiotherapy. Normal
used and the concentration achieved (assuming a uniform “B
tissue tolerance dose levels require us to limit the dose as indi-
distribution) are indicated in the figure.
cated achieving the response indicated by point “1.” Using
BNCT to locally enhance the effective dose to the tumor by an
increment A shifts the response to the level indicated by point
course of neutron radiotherapy. A detailed comparison “2.” The magnitude of the change in tumor control probability
between the quantitative predictions of our model cal- is indicated by I’.
culations and the experimental data shown in Figure 5
shows that the model predicts the measured enhanced
cell survival probability, and N is the number of cells in
cell killing to within a factor of 2-the model predicting
the tumor being treated. Eq. 2 assumes that the proba-
a somewhat larger effect than experimentally observed.
bilities of killing the individual tumor cells are indepen-
Given the basic simplicity of the model, this agreement
dent and that all cells must be killed to achieve tumor
is quite satisfying although if one wished to pursue it fur-
control. Using Eq. 2 we can then estimate the expected
ther, one could argue that this difference is consistent with
change in tumor control probability for a given change
the cells-that is, a lower concentration in the nucleus
in the cell survival probability. The resulting change de-
than in the cytoplasm, cell membrane, and surrounding
pends upon the value of the expected tumor control in
medium.
the absence of BNCT enhancement and results for un-
enhanced values of 0.2 and 0.5 are shown in Table 1, The
DISCUSSION range chosen for the expected change in cell survival in-
duced by BNCT was chosen from Figure 4. In the context
In summary, we have discussed the enhancement effect of this model the change in tumor control saturates for
of BNCT for our fast neutron radiotherapy beam in terms larger increases in cell killing. The Poisson model may
of a two component system-a “fast” and a “slow” com- well overestimate the actual change in tumor control but
ponent. We have measured these two components, mod- nevertheless, if an additional factor of lo-100 increased
eled the enhancement effect in terms of a generalized LQ cell killing can be achieved by BNCT enhancement, the
model, and performed experimental in vitro measure- change in tumor control should be clinically significant.
ments which indicate a satisfactory degree of agreement A major hurdle faced in using “pure” BNCT to kill
between theory and experiment. tumors relates to having to “tag” almost all of the tumor
How might we expect this change in cell survival prob- cells with “B. Otherwise, there is a high probability that
ability to impact tumor control? The point to remember the “untagged” cells will not be killed. In the case of using
is that dose response curves for both tumors and normal BNCT to enhance the effect of fast neutron cell killing,
tissue have an approximately sigmoidal shape as shown
in Figure 6. Using fast neutrons alone we are working on
the steep portion of the curve. A small change in cell Table 1. Estimated change in tumor control probability for a
given change in cell survival due to BNCT enhancement
killing due to BNCT can move us an appreciable distance based upon the Poisson model of Porter (27, 28)
up the response curve. The Poisson model as used by
Porter (27, 28) provides a simple way of estimating the Change in cell survival Change in tumor control
effect of this shift along the curve. This model relates a
s + 0.5 s 0.2 - 0.45
change in cell survival to an associated change in tumor 0.5 - 0.70
control via: s + 0.1 s 0.2 + 0.85
p = e+‘Js. 0.5 - 0.93
(Eq. 2) S+O.Ol s 0.2 + 0.98
0.5 + 0.99
In Eq. 2, P is the probability of tumor control, S is the
II40 I. J. Radiation Oncology 0 Biology 0 Physics Volume 28, Number 5, 1994

this criteria is less rigorous. There is no doubt that the maximize the number of slow neutrons in the beam which
enhancement effect is maximized if all tumor cells are could be simply accomplished by drastically degrading
“tagged” with “B but since the enhancement is statistical the “fast” component. What we want to do is to optimize
in nature, this only means that one does not move as far the relative tumor kill to that of the normal tissue damage.
along the dose response curve if some of the cells are not The optimal “mix” of fast and slow neutrons will depend
“tagged.” This effect is shown in Figure 7 which shows a both upon the differential uptake of l”B between the tumor
family of curves calculated assuming that only the per- and normal tissues as well as the relative location of the
centage of the tumor cells indicated on the lower right- tumor-that is, is it “deep” or “superficial.” It may well
hand axis take up “B and that the l”B tagged cells expe- be necessary to develop a library of different beams for
rience an additional factor of 10 in cell killing. As can be various clinical situations.
seen from this figure, the amount of the enhancement A second area of investigation relates to modeling the
depends both upon the probability of the unenhanced effect of the BNCT boost in clinical situations. The overall
tumor control and the number of tumor cells taking up increase in the physical dose is quite small compared to
the “B. For example, if only 50% of the tumor cells take its biological effect and so one cannot simply add the in-
up the “B and the unenhanced tumor control probability creased energy dose to the fast neutron dose. For the same
is 0.5, then the enhanced tumor control probability would increase in the BNCT physical dose, the radiobiological
be 0.683 instead of the 0.93 shown in Table 1 (which effect depends on the starting point on the dose response
assumes that all the tumor cells take up the i”B). This curve-that is, the location of point “1” in Figure 6. The
change is still large enough to be clinically relevant. net enhancement will need to be determined for both the
Applying BNCT to “boost” the therapeutic effectiveness tumor and the normal tissues in the treatment field and
of fast neutron radiotherapy has been proposed as well the resultant plotted as isoeffect contours rather than sim-
by workers at Essen (3 1, 36) and recently they have dem- ply isodose contours. Multiple treatment fields will be used
onstrated the effect in vivo using an implantable mela- and for clinical use, the computer modeling code must
noma model in mice (30). However, much work must be be fast enough for “real time” use-just as is the case for
done before clinical trials evaluating this technique can treatment planning codes used in today’s radiotherapy
be initiated. clinics.
One problem area relates to beam optimization. It is The need to develop better “B carriers is a common
important to remember that one does not simply want to problem for both “pure” BNCT and its use as a concom-
itant boost for fast neutron radiotherapy. The thermal
and epithermal beams currently under investigation for
“pure” BCNT are poorly penetrating and limit its appli-
cability to superficial tumors. A BNCT boost for fast neu-
trons can be applied to any tumor system where fast neu-
trons are used. This includes pelvic tumors such as pros-
0.9
tate cancer, intrathoracic tumors such as nonsmall-cell
0.9
lung cancer, as well as sarcomas located in various regions
0.6 0.8
of the body. Hence, it may be necessary to develop a dif-
0.7 0.7
ferent spectrum of “B carriers than are currently being
0.6 0.6
studied for “pure” BCNT. Currently fast neutron radio-
0.5 0.5
0.4
therapy is given on a fractionated basis- 12 or 16 fractions
0.4
0.3
over a 4 week period-rather than the single fraction most
0.3
0.2
commonly used in “pure” BNCT. In determining the op-
0.2
0.’
timal fractionation schedule when a BNCT boost is used,
0.1
<’
:0
the pharmacokinetic properties of the “B carriers will be
of critical importance. One can envision using different
agents at different time points in the course of therapy.
UNELNHANCED TUMOR :ING UP
CoNTFa-
For example. one might wish to use carriers based upon
hypoxic cell sensitizers (23) early in therapy when the
Fig. 7. Family of curves showing the magnitude of the BNCT
hypoxic cell fraction is greatest. For compounds that de-
enhanced tumor control probability as a function of the unen-
hanced tumor control probability and the percentage of tumor pend upon increased capillary permeability to penetrate
cells taking up the “B. An additional factor of 10 in cell killing the tumor, their point of use would be dependent upon
was assumed for the BNCT effect and the Poisson model was the particular tumor treated. For example, if one were
used to estimate the magnitude ofthe enhancement. Each curve treating a malignant glioma, then such agents would be
represents a specific fraction of tumor cells taking up ‘“B as
used early in the course of therapy when one is dependent
indicated on the lower right axis. The unenhanced tumor control
probability (i.e., in the absence of “B) is shown on the lower upon an intact blood-brain barrier to keep the “B out of
left axis. the normal brain parenchyma. However, if one were
 BNCT in fast neutron radiotherapy 0 G. E. LARAMORE elal. 1141

treating a sarcoma of the extremity, then it might be best Local-regional tumor control is, of course, only part
to use such an agent later in the course of therapy when of the problem of improving patient survival but for pri-
increased capillary permeability would allow the agent to mary brain tumors and cancers of the head and neck,
more readily penetrate the tumor tissue. The timing be- inability to achieve it is the major cause of patient death
tween the administration of the l”B carrier and the delivery (22, 33). For other tumors where distant metastases are a
of the fast neutron radiation will also be important in more important component of “first failure,” improved
maximizing the therapeutic gain factor. Just as in con- local-regional control will still be an important step for-
ventional BNCT it will be important to use compounds ward. BNCT enhancement of fast neutron radiotherapy
that produce a high differential in “B concentration be- may be an important avenue to achieving improved sur-
tween tumor and adjacent normal tissues at the time the viva1 for selected patient groups and is well worth further
radiation is given. investigation.

REFERENCES

1. Asbury, A. K.; Ojemann, R.; Nielsen, S. L.; Sweet, W. H. 14. Hatanaka, H.; Sane, K. A revised boron neutron capture
Neuropathologic study of fourteen cases of malignant brain therapy for malignant brain tumors. I. J. Neurol. 204:309-
tumors treated by boron- 10 slow neutron capture therapy. 332; 1973.
J. Neuropathol. Exp. Neural. 31:278-303; 1972. 15. Kobayashi, T.; Kanda K. Analytical calculation of boron-
2. Brahme, A.; Eenmaa, J.; Lindbaeck S.; Montelius, A.; 10 dosage in cell nucleus for neutron capture therapy. Ra-
Wootton, P. Neutron beam characteristics from 50 MeV diat. Res. 9 1:77-94; 1982.
protons on beryllium using a continuously variable multi- 16. Kruger, P. G. Some biological effects of nuclear disintegra-
leaf collimator. Radiother. Oncol. 1:65-76; 1983. tion products on neoplastic tissue. Proc. Natl. Acad. Sci.
3 Brownell, G. L.; Murray, B. W.; Sweet, W. H.; Wellum, 26:181-192; 1940.
G. R.; Soloway, A. H. A reassessment of neutron capture 17. Laramore, G. E. Injury to the central nervous system after
therapy in the treatment of cerebral gliomas. Seventh Na- high LET radiation. In: Gutin, P. H., Liebel, S. A., Sheline,
tional Cancer Conference Proceedings. Philadelphia: Lip- G. E., eds. Radiation injury to the nervous system. New
pincott; 1973827-837. York: Raven Press; 199 1:34 l-360.
4. DeBenedetti, S. Nuclear Interactions. New York: John 18. Laramore, G. E.; Austin-Seymour, M. Fast neutron radio-
Wiley; 1964:338-344. therapy in relation to the radiation sensitivity of human
5. Dewit, L.; Moss, R.; Gabel, D. New developments in neutron organ systems. In: Altman K. I., Lett, J.. eds. Relative ra-
capture therapy. Eur. J. Cancer 26:9 12-9 14; 1990. diosensitivities of human organ systems, III. Advances in
6. Fairchild, R. G.; Bond, V. P. Current status of “B-neutron radiation biology, Vol. 15. Orlando: Academic Press; 1992:
capture therapy: Enhancement of tumor dose via beam fil- 153-193.
tration and dose rate, and the effects of these parameters on 19. Laramore, G. E.; Griffeth, J. T.; Boespflug, M.; Pelton,
minimum boron content: A theoretical evaluation. Int. J. J. G.; Griffin, T.; Griffin, B. R.; Russell, K. J.; Koh, W,;
Radiat. Oncol. Biol. Phys. 11:831-840; 1985. Parker, R. G.; Davis, L. W. Fast neutron radiotherapy for
7. Farr, L. E.; Sweet, W. H.; Robertson, J. S.; Foster, C. G.; sarcomas of soft tissue, bone, and cartilage. Am. J. Clin.
Locksley, H. B.; Southerland, D. L.; Mendelsohn, M. L.; Oncol. (CCT) 12:320-326; 1989.
Stickley, E. E. Neutron capture therapy with boron in the
20. Laramore, G. E.; Krall, J. M.; Griffin, T. W.; Duncan, W.;
treatment of glioblastoma multiforme. Am. J. Roentgenol.
Richter, M. P.; Saroja, K. R.: Maor, M. H.; Davis, L. W.
71:279-291; 1954.
Neutron vs. photon irradiation for unresectable salivary
8. Gabel, D.; Foster, S.; Fairchild, R. G. The Monte Carlo
gland tumors: Final report of an RTOG-MRC randomized
simulation of the biological effect of the “B(n,a)‘Li reaction
clinical trial. Int. J. Radiat. Oncol. Biol. Phys. 27:235-
in cells and tissue and its implication for boron neutron
240;1993.
capture therapy. Radiat. Res. 111: 14-25; 1987.
21 Laramore, G. E.; Krall, J. M.; Thomas, F. J.; Russell,
9. Godwin, J. T.; Farr, L. E.; Sweet, W. H.; Robertson, J. S.
K. J.; Maor, M. H.; Hendrickson, F. R.; Martz, K. L.; Griffin,
Pathological study of eight patients with glioblastoma mul-
T. W.; Davis, L. W. Fast neutron radiotherapy for locally
tiforme treated by neutron capture therapy using boron 10.
advanced prostate cancer: Final report of an RTOG ran-
Cancer (Phila.) 8:601-615: 1955.
domized clinical trial. Am. J. Clin. Oncol. (CCT) 16:164-
10. Griffin, T. W.; Martz, K. D.; Laramore, G. E.; Thomas,
167; 1993.
F. J.; Maor, M. H.; Hendrickson, F. R.; Richter, M. P.;
Davis, L. W. High energy (42-66 MeV) reactions) fast neu- 22.. Leibel, S. A.: Ling, C. C.; Kutcher, G. J.; Mohan, R.; Cordon-
tron dose optimization studies in the head and neck, thorax, Cordo, C.; Fuks, Z. The biological basis for conformal three-
upper abdomen, pelvis, and extremities. Radiother. Oncol. demensional radiation therapy. Int. J. Radiat. Oncol. Biol.
19:307-316; 1990. Phys. 21:805-811; 1991.
Il. Griffin, T. W.; Pajak, T. F.; Laramore. G. E.; Duncan, W.; 23 Livesey, J. C.; Wiens, L. W.; Wilbur, D. S.; Hamlin, D. H.;
Richter, M. P.; Hendrickson, F. R.; Maor, M. H. Neutron Laramore, G. E. Hypoxia-selective cellular toxicity of car-
vs. photon irradiation of inoperable salivary gland tumors: boranyl-nitroimidazole. Proceedings of the Fifth lnterna-
Results of an RTOG-MRC cooperative randomized study. tional Symposium on Neutron Capture Therapy for Cancer.
Int. J. Radiat. Oncol. Biol. Phys. 15:1085-1090; 1988. New York: Plenum Press; 1993 (In press).
12. Hatanaka, H. A revised boron neutron capture therapy for 24. Lecher, G. L. Biological effects and therapeutic possibilities
malignant brain tumors. II. J. Neurol. 209:8 l-94; 1975. of neutrons. Am. J. Roentgenol. 36: I- 13: 1936.
13. Hatanaka H. Boron neutron capture therapy for tumors. 25. Mijnheer, B. J.; Wootton, P.; Williams, J. R.; Eenmaa, J.;
Niigata, Japan: Nishimura Press; 1986. Parnell, D. J. Uniformity in dosimetry protocols for the
1142 I. J. Radiation Oncology 0 Biology 0 Physics Volume 28. Number 5, I994

therapeutic applications of fast neutron beams. Med. Phys. Symposium on Neutron Capture Therapy for Cancer. New
14:1020-1026; 1987. York: Plenum Press; 1993 (In press).
26. Mishima, Y.; Ichihashi, M.; Hatta, S.; Honda, C.; Yama- 3 1. Sauerwein, W.; Ziegler, W.; Olthoff, K.; Streffer, C.; Rassow,
mura, K.; Nakagawa, T.; Obara, H.; Shirakawa, J.; Hirast- J.; Sack, H. Neutron capture therapy using a fast neutron
suka, J.; Taniyama, K.; Tanaka, C.; Kanda, K.; Kobayashi. beam: Clinical considerations and physical aspects. Strah-
T.; Sato, T.; Ishida, M. R.; Ujeno, Y.; Takahashi, M.; Abe, lenther. Onkol. 165:208-210; 1989.
M.; Nozaki, T.; Aizawa, 0.; Matsumoto, T.; Sato, T.; Kar- 32. Stone, R. S. Neutron therapy and specific ionization. Am.
ashima, H.; Yoshimo, K.; Fukuda, H. First human clinical J. Roentgenol. 59:771-785; 1940.
trial of melanoma capture: Diagnosis and therapy. Strah- 33. Suit, H. D.; Westgate, S. J. Impact of improved local control
lenther. Onkol. 165:25 l-254; 1989. on survival. Int. J. Radiat. Oncol. Biol. Phys. 12:453-458;
27. Porter, E. H. The statistics of dose/cure relationships for 1986.
irradiated tumors. Part I. Br. J. Radiol. 53:210-227; 1980. 34. Waterman, F. M.; Kuchnir, F. T.; Skaggs, L. S.; Bewley,
28. Porter, E. H. The statistics of dose/cure relationships for D. K.; Page, B. C.; Attix, F. H. The use of l”B to enhance
irradiated tumors. Part II. Br. J. Radiol. 53:336-345; 1980. the tumour dose in fast-neutron therapy. Phys. Med. Biol.
29. Russell, K. J.; Laramore, G. E.; Krall, J. M.; Thomas, 23:592-602; 1978.
F. J.; Maor, M. H.; Hendrickson, F. R.; Krieger, J. N.; Grif- 35. Zahl, P. A.; Cooper, F. S.; Dunning, J. R. Some in viva
fin, T. W. Eight years experience with neutron radiotherapy effects of localized nuclear disintegration products on a
in the treatment of stages C and D prostate cancer: Updated transplantable mouse sarcoma. Proc. Nat. Acad. Sci. 26:
results of the RTOG 7704 randomized clinical trial. Prostate 589-598; 1940.
11:183-193; 1987. 36. Ziegler, W.; Sauerwein, W.; Streffer, C. Fast neutrons from
30. Sauerwein, W.; Heselmann, I.; Sack, H. Neutron capture the Essen cyclotron can be used successfully for neutron
reactions in fast neutron therapy can improve local tumor capture experiments in vitro. Strahlenther. Onkol. 165:2 I O-
control in mice. Proceedings of the Fifth International 212; 1989.