Journal of Neuro-Oncology 33: 19–26, 1997.

 1997 Kluwer Academic Publishers. Printed in the Netherlands.

Early history of development of boron neutron capture therapy of tumors

William H. Sweet
Neurosurgical Service, Massachusetts General Hospital, Boston, MA 02114, USA

Key words: history, neutron capture treatment, boron, lithium, glioma, melanoma

Summary

The stable isotope 10B has a peculiarly marked avidity to capture slow neutrons whereupon it disintegrates
into a lithium and a helium atom. These give up the 2.4MeV of disintegration energy which they share within 5
and 9µm of the 10B atom respectively. This means that the cell closest to the 10B atom bears the brunt of its
atomic explosion. The objective of the tumor therapist is to find a carrier molecule for the boron atom which
will concentrate in the tumor. Although a number of investigators saw the peculiar advantage of this selective
tactic to achieve destruction of a species of unwanted cells, no success in animal studies was achieved until
1950. Sweet and colleagues found that the capillary blood-brain barrier keeps many substances out of the
normal brain but that the gliomas had much less of such a barrier. He, Brownell, Soloway and Hatanaka in
Boston together with Farr, Godwin, Robertson, Stickley, Konikowski and others at the Brookhaven National
Laboratory worked partially in collaboration and partly independently. We irradiated at 3 nuclear reactors
several series of glioma patients with no long-term remission, much less a cure being achieved. Hatanaka on
his return to Japan kept BNCT alive by treating a total of 140 patients with various brain tumors. Beginning in
1972, Mishima and colleagues have achieved useful concentrations of 10B-borono-phenylalanine, an analogue
of the melanin precursor tyrosine, for BNCT of melanomas.

My historical account of the evaluation of the earli-
er thinking and the observations on neutron cap-
ture therapy follows recent comprehensive contri-
butions by Slatkin [1] and Hawthorne [2]. For my
brief study I shall concentrate first on the earliest
stages of biological knowledge pertinent to neutron
capture therapy with brief reference to many mea-
gerly explored alleys trod by investigators.
An electrically neutral elementary particle, the
neutron, with a mass 0.782 MeV, more than that of
the proton, was first identified in 1932 by Chadwick
at Cambridge University’s Cavendish Laboratory
[3]. The fast neutrons produced by bombarding ber-
yllium with alpha particles had energies up to
4.5MeV. Fermi discovered that neutrons react most
efficiently with a number of elements after they are
slowed by passage through a hydrogen-rich sub-
stance such as paraffin [4]. Goldhaber, collaborat-
ing with Chadwick [5], Taylor [6] and Burcham [7]
showed that slow-neutron bombardment of specific
stable isotopes of boron, lithium and nitrogen yield-
ed charged-particle tracks in photographic plates.
The tracks from boron disintegration were short
and straight, consistent with formation of 2 particles
travelling in opposite trajectories. Their average
distance of travel in this photographic gelatin was
approximately 7.6µm. Subsequently the complete
reaction was shown to be:
10
B + 1nth ➙ [11B] ➙ 7Li + 4He + 0.48 MeV + 2.31 MeV 94%
➘ 7Li + 4He + 2.79 MeV. 6%
By 1936 Locher had published a comprehensive
theoretical account of biological effects and ther-
apeutic possibilities of these particles [8]. It makes
fascinating reading, decades later, to see what facts

Please indicate author’s corrections in blue, setting errors in red

121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z

20
relevant to biological studies needed to be known if
one were to embark on such studies. Locher point-
ed out that the first requirement, a source of ade-
quate flux of neutrons, would be difficult to achieve.
Unlike alpha particles, which are emitted spontane-
ously by 23 radionuclides, as far as he knew, neu-
trons were ‘not spontaneously emitted by any
known element’. He went on to say ‘If neutron-
emitting radioactive elements ever existed in the
earth’s crust, they must have decayed and had no
long-lived progenitors’. Unknown to Locher in
1936 were nuclear reactors and ultra-long-lived ura-
nium and thorium isotopes.
However, Locher gave a detailed description of
the many biological effects expected from irradia-
tion by neutrons. In the field of medical research he
cited ‘‘the possibility of destroying or weakening
cancerous cells, by the general or selective absorp-
tion of neutrons by these cells. In particular there
exist the possibilities of introducing small quantities
of neutron absorbers into the regions where it is de-
sired to liberate ionizing energy. A simple illustra-
tion would be the injection of a soluble non-toxic
compound of boron, lithium, gadolinium or gold in-
to a superficial cancer, followed by bombardment
with slow neutrons’’. He gave 4 pages of suggestions
on how to generate an adequate flux of slow neu-
trons, an objective achieved only recently. The need
for exact quantification of all relevant physical par-
ameters may be illustrated by the capture cross sec-
tion values in barns (1 barn = 10−24 cm2) for slow neu-
trons of the naturally occurring mixture of stable
isotopes of boron (atomic weight 10.82) being 760
instead of 3838 for purified 10B.
Kruger and Goldhaber began experiments with
mouse tumors in 1938. 9 They noted that the massive
energies of fission of 0.8 MeV for the 7Li and
1.4MeV for the alpha particle are dissipated in ap-
proximately 5 and 9 microns respectively in a single
linear path in tissue. They emphasized that these
parameters yield a far more intense ionization than
occurs by a proton recoiling from a collision with a
fast neutron. They recognized the major advantage,
in confining nearly all of the radiation to the site of
the boron atoms, is the potential that this system
offered in protecting nearby normal tissues. Kruger
worked with 3 murine tumors: a sarcoma, a mam-
Please indicate author’s corrections in blue, setting errors in red
121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z
mary carcinoma and a lymphoma. Tumor speci-
mens were removed and the size of these that con-
sistently produced regrowth upon transplantation
was established. Boric acid was injected into many
tumor specimens before reimplantation. Controls
included specimens injected with the boron but not
radiated before reimplantation. Slow neutron dos-
es to the reimplanted tumors containing H3BO3
were established as being much lower to prevent tu-
mor regrowth than were required for X-rays or fast
neutrons.
Zahl, Cooper and Dunning carried the studies a
step further by developing a transplantable mouse
sarcoma that grew consistently to kill the animal in
3 to 4 weeks after implantation in the mouse’s axilla
[10]. By direct injection into and around the tumor
of various preparations of boron or lithium fol-
lowed by many hours of irradiation with their feeble
neutron source, they did not succeed in achieving a
sufficiently homogenous distribution of the boron
or lithium within the tumor to destroy it without
killing the mouse. A year later Zahl and Cooper re-
ported on their efforts to modify acid dyes which, on
i.v. injection, would concentrate in tumors [11].
They substituted lithium for sodium in the dye mol-
ecule. They found lithium carmine did dye most
heavily the peripheral zones of both rat and murine
sarcoma. They pointed out that this zone is assumed
to contain the most frequently dividing malignant
cells and hence is the most advantageous position
for the lithium dye. However, they found that, once
injected, the lithium and the dye moiety separated.
Within 24 hours after injection of Li-carmine, lithi-
um was no longer detectable spectroscopically in
the animal, whereas the dye’s color persisted in the
tissue for many days. Many years later, 15 vital dyes
were screened in tumor-bearing mice at Brook-
haven by Sinex et al. in 1953 [12].
McClintock and Friedman described the use of
antibodies to localize uranium to a specific tissue
[13]. McClintock’s untimely death precluded publi-
cation of the method. Knock at first continued the
use of a bentonite-uranium [14] complex, and then
the antibody approach of McClintock. She succeed-
ed in depositing over 5mg/ml of uranium in an ani-
mal’s calf by perfusing uranium-antibody complex-
es through the femoral artery and vein in the groin.

Although she described many persisting problems
and her proposed solution, the work was not contin-
ued.
In 1948 Tobias and others pointed out that the use
of atomic fission in vivo had potential applications
in radiation therapy if fissionable elements could be
incorporated into compounds that localize at spe-
cific sites in the body [15]. They studied the distribu-
tion of uranium in various murine tissues after in-
travenous injection and assessed acute lethal effects
in an experiment designed to evaluate the biolog-
ical effectiveness of uranium fission. This compre-
hensive effort involved 7 different animal groups
together with appropriate controls. A variety of
toxic effects of the uranium were carefully delineat-
ed. Most startling was the energy liberated per atom
of 235U at disintegration, 159 MeV, which makes bo-
ron fission look feeble by comparision. The 235U val-
ue includes the radioactivity of the fission products
[16]. Our group, 10 years later, undertook the study
of uranium in man. These studies were carried out,
not only with a view to assessing the potential use of
uranium in brain tumor therapy, but also to provide
data on human toxicology of uranium, thereby de-
lineating tolerance levels for uranium workers [17,
18]. This was necessary because of the extreme vari-
ation between animal species, e.g. for its weight, the
mouse tolerates 100 to 200 times more uranium than
the rabbit. The renal toxicity of hexavalent uranium
given intravenously in man is demonstrated by the
appearance of catalasuria and albuminuria. The
minimal dose necessary to produce this effect is of
the order of 0.1mg of hexavalent uranium/kg body
weight. Two substantial articles, Struxness et al. [17]
and Luessenhop et al. [18], combined the work at
Oak Ridge National Laboratory with that of the
neurosurgical service at the Massachusetts General
Hospital. Their results demonstrated that uranium,
in the forms used, was too toxic for human use in
neutron capture therapy.
The foregoing work established the special vir-
tues of boron and lithium for neutron capture ther-
apy of tumors. However, it did not establish any
general principle or criteria by which the boron or
lithium moieties would be taken up selectively by
the tumor cell. Our approach to a therapeutic trial
of boron neutron capture therapy (BNCT) began
Please indicate author’s corrections in blue, setting errors in red
121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z
21
by trying to get a useful concentration of a boron-
containing agent into the brain tumor. Only subse-
quently would the second component, namely the
thermal neutrons, be in order to generate a cytotox-
ic effect. The principle we have sought to apply is
valid only for the normal CNS whose blood-brain
barrier (BBB) is characterized by the tight junc-
tions of the endothelial cells of its capillaries. The
objective was for the intact BBB to prevent normal
areas of the brain from taking up the neutron cap-
turing isotope, in much the same way that it has re-
stricted the entry of other specific isotopes.
In 1949 Selverstone, Solomon, Robinson and I re-
ported using Geiger-Müller counters incorporated
into metal tubes 2 or 3mm in diameter to detect the
beta radiation from 32PO4 [19, 20] in tumor present
in human brain. The phosphate ion given intrave-
nously concentrated to a spectacular degree in all 8
tumor types that we studied, including the relatively
slow growing types: astrocytoma, ependymoma,
and oligodendroglioma. Of 14 glioblastomas, 7 had
tumor:normal brain ratios of 5.8 to 9.7 and in the
remaining 7, the values ranged from 10.8 to 73.4.
The median ratio in the glioblastoma cases with the
highest concentrations was 22.6. Moreover, these
ratios tended to persist for more than 24 hours. The
fact that normal brain controls so rigorously exclud-
ed the entry into itself of such an important ion for
energy production and DNA repair was astounding
to us! The capacity of the brain to keep out phos-
phate and of the brain tumor to actively accrete and
retain it provides a meaningful example that our
quest for the best boron carrier still remains to be
achieved.
The thought of using the electrons from the same
phosphorus radioisotope to kill the tumor cells was
negated by the high uptake of the phosphate in rap-
idly metabolizing normal tissues such as gastroin-
testinal mucosa, bone marrow, and especially liver,
which takes up much more 32P than does glioblasto-
ma [21, 22]. The closed junction capillary endothe-
lial barrier between the blood stream and normal
tissue exists only for the central nervous system.
There is no capillary barrier that protects other nor-
mal tissues from too rapid an exchange between it
and the blood stream. The 32PO4 behavior clearly
pointed to the requirement for a bipartite mecha-
22
nism: (1) the need for a relatively innocuous sub-
stance that avidly binds to many rapidly metaboliz-
ing tissues but especially malignant brain tumor
cells. This must be followed by (2) an agent or beam
directed only at the brain that converts that innoc-
uous substance to a cytocidal entity or entities capa-
ble of destroying the tumor.
The 1950 paper of Conger and Giles at Oak Ridge
reported that the trace amounts of boron normally
present in lily bulbs were responsible for most of the
radiation changes in the plants following exposure
to slow neutrons [23]. This enunciated the biolog-
ical facts clearly and simply and led directly to our
conclusion, namely, to see if the normal brain would
exclude enough of a boron compound so that it
would withstand a dose of slow neutrons while the
more readily permeable tumor cells would suc-
cumb.
In collaboration with Brownell [24] we computed
to a first approximation the radiation dosage arising
from neutron capture of all of the isotopes in nor-
mal and neoplastic brain according to the following
simple formula:
Radiation Dose = F × N × a × E × A
F = Flux = number of thermal neutrons/cm2/sec
N = Number of atoms of an isotope/unit volume
σa = Capture or absorption cross section for ther-
mal neutrons
E = Energy in MeV generated by fission at the
time of capture reaction
A = Fraction of this energy absorbed in tissue
The computations revealed that if we had a concen-
tration of 10B in tumor of 50 µg/g while the normal
gray matter achieved only 15 µg (concentration ra-
tio 3.33), the tumor would receive 3 times the radi-
ation dosage as the normal brain. At these concen-
tration levels, 10B largely determines the radiation
effect. With the first compound studied, borax
(Na2B4O7⋅1OH2O), 15–20 µg of boron/g of brain was
not toxic. Tumor:brain boron ratios of 3 to 28 were
obtained but the duration of this ratio was only for
short time intervals. We decided we were ready to
use the nuclear reactor as our source of the second
relatively innocuous factor, the thermal neutrons.
The Brookhaven National Laboratory’s first ma-
Please indicate author’s corrections in blue, setting errors in red
121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z
jor investigative device was its 20 megawatt nuclear
research reactor commissioned in 1950 and capable
of easily providing a useful thermal neutron beam.
We presented our data to Dr. Lee Farr, the Chair-
man of the Brookhaven Medical Department, and
found that he was already searching for a suitable
carrier for a capture element. So we happily joined
forces. Critically aided by the massive resources
and eager physicists at Brookhaven, we treated 10
glioblastoma patients who had had a grossly total
resection of their tumors at the Massachusetts Gen-
eral Hospital. Verbal consent was on behalf of each
patient prior to their participation in these early in-
vestigational trials using BNCT to treat their termi-
nal disease.
A portion of the shielding atop the reactor was
removed to permit placing the lateral aspect of the
patient’s intact scalp and skull at the specially de-
signed portal. To prevent scalp damage, we tied off
the external carotid arteries and covered the entire
scalp with tight elastic bandages in an attempt to
prevent boron-containing blood from entering the
scalp. These tactics however, did not prevent the de-
velopment of several large radiation erosions of the
scalp. Five patients received a single radiation dose,
and the remaining 5 were given the treatment in 2 to
4 fractions. Although there were no life-threatening
complications of the therapy, all of the patients died
from 6 to 21 weeks after the first session of neutron
capture therapy, which was usually the case during
the 1950s for glioblastoma patients treated by any
means [25]. Post mortem studies done in 6 of the
patients showed abundant viable tumor [26, 27].
Their painful scalp lesions together with the inade-
quacy of the radiation dose led us to attempt to de-
liver the thermal neutron beam directly to the
grossly normal but microscopically, tumor-infiltrat-
ed brain. The Rockefeller Foundation made this ap-
proach possible with a $500,000 gift to the Massa-
chusetts Institute of Technology to provide addi-
tional features to a nuclear reactor that was then be-
ing constructed. Included was a surgical operating
room immediately beneath the reactor core. This
permitted us to turn down again the scalp, bone and
dural flaps used in the prior removal of gross tumor.
At reopening, the cerebrospinal fluid replacing tu-
mor was also drained away to give maximally un-

impeded access of the thermal neutrons through
sterile air to the tumor-infiltrated brain.
At the Massachusetts General Hospital a contin-
uing systematic effort to find a suitable non-toxic
chemical containing adequate boron was mounted
in the 1950’s by Soloway and his colleagues [28–30].
He maintained a steady stream of improved boron
compounds beginning in 1958 with a study of the
features a molecule should have in order to be re-
stricted in entering normal brain and/or accelerated
in entering brain tumor from the blood stream. He
studied a series of mainly monosubstituted deriv-
atives of phenylboronic acid to learn the ratio of tu-
mor:brain content of each. This ratio was compared
to the partition ratio of the compound between the
lipid solvent, benzene, and water; the penetration of
the brain by a drug being a function of its lipid solu-
bility. He showed with numerous examples that the
higher the tumor : brain concentration ratio the
higher the partition coefficient in the aqueous rath-
er than in the benzene phase. Conversely, increased
entry into normal brain is associated with toxicity
and a higher partition coefficient in the benzene
phase. Among the monosubstituted examples of
phenylboronic acid that he studied, the one show-
ing the highest tumor : brain ratio, was p-carboxy-
phenylboronic acid. After the intraperitoneal injec-
tion of this compound into mice with subcutaneous-
ly-transplanted gliomas, their glioma-brain ratios
varied from 2.5 to 9 when measured from 1/4 hr to
3 hr after injection. This compound was used as the
10
B carrier for 16 of the 18 patients treated by BNCT
at the MIT reactor. Subsequently, sodium perhy-
drodecaborate (Na2B10H10) proved to be the most
soluble, chemically stable and biologically inert of
all of the boron carriers that Soloway had tested at
that juncture. It turned out to be equally non-toxic
when injected into the carotid artery. This site of in-
jection yielded a higher level of compound in the
brain tumor than did the i.v. route. Hence the
Na2B10H10 was chosen as the boron carrier for the
last 2 patients and the carotid, the site of its injec-
tion. Prior to their involvement in this clinical
BNCT trial, verbal consent was obtained on behalf
of each patient.
The 18 patients treated at MIT died in from 10
days to 11 1/2 months after radiation and the de-
Please indicate author’s corrections in blue, setting errors in red
121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z
23
tailed studies of the 14 post mortem brains have
been published [31]. In every patient the cause of
death was cerebral. Extensive radiation necrosis of
the brain was induced in 9 cases, in 2 cases, only re-
current tumor was seen and in one patient, there
was extensive radiation necrosis and tumor. In only
1 of the 11 glioblastoma patients were islands of glio-
blastoma no longer seen. In this patient the total
neutron fluence at the surface of the brain was only
1.1× 1013/cm2; this was 1/3 of the highest dose given in
this series. The radiation time of 45min was less
than 1/2 the longest time and the dose of 25mg/kg of
10
B was slightly less than the 30mg/kg given to 8 of
the patients. We could only conclude that this tumor
was less radioresistant than most of the series. The
crucial data of boron levels in blood, normal brain
and tumor were not measured.
The radiation necrosis in nearly all of the patients
was characterized by thickening of the blood vessel
walls due to proliferation and enlargement of all of
their cells resulting in vascular occlusion. A ‘coag-
ulation necrosis’ of the devitalized tissue rather
than the usual liquefactive chain of events took
place with especially striking blood vessel affliction.
The inference was that the fissioning boron in the
blood stream induced the vascular changes. The ar-
eas of the brain showing coagulative necrosis ap-
pear not to be invaded by any tumor but are ren-
dered ischemic by the proliferative reaction of the
various components of the blood vessels due to fis-
sion products. From these results, it is clear that the
dosimetry must include the boron in the blood ves-
sels both those in completely normal brain and
those invaded to various degrees by tumor. Sophis-
ticated attention has been given to this problem by
Kitao [32] of the National Institute of Radiological
Sciences in Japan who was stimulated to do this by
Hatanaka and by Rydin, Deutsch and Murray at
MIT [33]. Using different methods, the 2 centers
compared relative boron doses to a 9µm diameter
cylinder and a 12.5 µm diameter sphere. The MIT
group calculated 0.67 and 0.70 respectively; Kitao’s
results were 0.67 and 0.66. A rule of thumb [34] is
that the radiation dose to the blood vessel wall is
comprised of 1/3rd from the intraluminal concen-
tration of boron and 2/3rds from the extravascular
boron level. More recently, Slatkin [1] used this rule
24
to estimate the radiation doses to capillary endo-
thelium in normal brain in 17 patients treated by
Farr, Yamamoto et al. at Brookhaven from 1959
through 1961 using 10B-enriched sodium pentabo-
rate as the capture agent.
During this entire period our quantitative deter-
minations of boron by colorimetry required 3–4
hours. Consequently, we determined the concen-
tration levels after a modest dose of the boron carri-
er at the time of the first craniotomy for tumor re-
moval. Then we waited at least 3 weeks for reconsti-
tution of the BBB before giving the treatment dose
of the boron and the neutron radiation. We as-
sumed that we could extrapolate from the curve at
the lower dose to calculate the probable blood level
at the higher dose at the relevant time. In our last
patient, the de facto concentration as measured
hours later from the blood sample drawn at the
treatment time was for the first time much higher
than our initial calculation. This led to cerebral ede-
ma and death 10 days after the treatment. We decid-
ed that we could conduct no further clinical trials
until we had boron compounds, which not only at-
tained suitable tumor:brain concentration differen-
tials but whose boron blood levels at the time of ir-
radiation would not generate vascular radiation ne-
crosis. Thus, one requirement was to assay blood
levels of boron within a few minutes of taking the
specimen.
For this latter prerequisite, we had to wait 24
years until Ralph Fairchild and his colleagues at
Brookhaven came up with an essentially perfect so-
lution to the problem [35]. They measured the con-
centration of fissionable 10B, which is only about
20% of the naturally occurring mixture of 10B and
11
B. The method utilizes the 478 keV gamma ray that
arises in 94% of the disintegrations of 10B, occurring
immediately from the capture reaction. The ther-
mal neutrons needed to run such determinations
are of course obtainable directly from the reactor
used for therapy.
The disappointing results of our MIT series also
spurred Soloway and Hatanaka to intensify their ef-
forts. By 1967 they had screened more than 150 bo-
ron compounds via their murine subcutaneously
transplanted ependymoblastoma and had found
that chemical inertness was less desirable than a
Please indicate author’s corrections in blue, setting errors in red
121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z
moiety that could attach somewhere to a tumor cell
and achieve tumor:blood differentials that were
greater than one [36]. The sulfhydryl group cova-
lently bound in the anion B12H11SH2− as the sodium
salt was the first compound to demonstrate such
properties. Tumor:blood boron ratios ranged from
1.4 to 20.0 (the higher values related to the initial
material provided by E.I. DuPont and contained
oxidized components). Often the boron concentra-
tion ini the normal murine brain was immeasurable
even if the entire brain was the sample. The
B12H11SH2− anion (BSH) or its metabolic products
were found at least partially to be covalently bound
to blood serum albumin. Conceivably, plasmaphe-
resis may be used to reduce the level of boron in the
blood stream. BSH is sufficiently toxic so that rapid
intravenous injections even of isotonic solutions
can be hazardous. However, slow i.v. injection of
240mg of boron/kg over 5 days, at a rate of 40mg/
kg/day, exerted no toxic effects. BSH per se, how-
ever, is unstable and readily oxidizes to more toxic
substances. The chemists at Shionogi in Japan as
well as Tolpin, Wellum and Berley at the Massachu-
setts General Hospital [37] succeeded in producing
a stable form of BSH for human use. Improvements
in the purification and stabilization of the BSH
were achieved by Hatanaka and his colleagues at
Shionogi permitting use of this compound as the bo-
ron carrier for the series of 140 patients [38].
After 3 years at the Massachusetts General Hos-
pital, Hatanaka returned in 1967 to the neurosurgi-
cal service at Tokyo University. Its chief, Professor
Keiji Sano, provided Dr. Hatanaka with the neces-
sary support to overcome many obstacles and treat
his first brain tumor patient in Japan by August
1968. In his preface to the second of Hatanaka’s
books, Sano notes that of Hatanaka’s 90 cases of
malignant gliomas, 40 were treated by BNCT and
the other 50 by a multimodality combination of
fractionated photon-radiotherapy and chemother-
apy [39]. The BNCT group, although 10 years older,
had a 5-year survival rate, 4 times that of his multi-
modality group. This achievement occurred despite
the availability to him of only 100kW reactors, in
contrast with the multi-megawatt reactors at
Brookhaven and MIT. From the beginning it was
clear to Hatanaka that tumors with components

deeper than 6cm probably would receive an inade-
quate dose of thermal neutrons due to the beam’s
limited tissue penetration. Hence, he was happy to
find a 5-year survival rate of 58% for patients with
glioblastoma multiforme having relatively superfi-
cial tumors. This value is strikingly better than the
4.7% 5 year survival figure recorded by Charles
Wilson in 1991 for his 449 patients with a diagnosis
of glioblastoma at the first operation. They were
treated by radical surgery plus radiation and che-
motherapy [40]. It is a tribute to Hatanaka that he
finally had many clinicians examining BNCT as a
potential modality for the treatment of malignant
brain tumors.
Yutaka Mishima and collaborators began in 1972
the clinical development of Snyder et al.’s [41] con-
cept involving the tyrosine analogue, boronophe-
nylalanine, as a boron carrier for neutron capture
therapy. This compound was initially screened by us
in brain tumor-baring animals [29]. Mishima stated,
‘‘Since 1972, I have developed the idea to selectively
kill melanoma (Mm) cells by coupling the high en-
ergy releasing system, 10B (n, α) 7Li reaction, with
the 10B-dopa (melanin substrate) analogue 10B1-L-p-
boronophenylalanine (10B1-BPA). After the synthe-
sis of 10B1-L-BPA fructose complex, exhaustive in
vitro and in vivo radiological studies on its en-
hanced killing effect were done to develop optimal
Mm NCT, and we have successfully treated 16 hu-
man Mm cases to date.’’ [42].
Fairchild, Glass, Coderre and Joel have pursued
BNCT studies at Brookhaven National Laboratory
with BPA, showing that this compound accumulat-
ed in non-melanotic tumors such as a rat gliosarco-
ma. Building on the expertise of the Japanese chem-
ists, who discovered how to solubilize BPA using
fructose, a team of investigators from Brookhaven
and the Beth Israel Hospital in New York under-
took a clinical BNCT trial for treating glioblasto-
mas using BPA as the capture agent, beginning in
September 1994. Their results are still being await-
ed but in contrast with the earlier trials in the Unit-
ed States, there is no evidence now of any radiation
morbidity or mortality stemming from the proce-
dure.
In the future, we can anticipate better neutron
beams, more effective tumor targeting agents and
Please indicate author’s corrections in blue, setting errors in red
121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z
25
the possibility of using BNCT in the treatment of
other neoplasms for which tumor control remains
an unachieved objective. For the present however,
the emphasis must be focused on malignant brain
tumors in view of the dismal prognosis for patients
with glioblastoma multiforme and the need to
prove unequivocally, whether or not BNCT has a
useful role to play in its treatment.
Acknowledgements
The author wishes to express his appreciation to the
Neuroresearch Foundation for its support during
the preparation of this manuscript.
References
1. Slatkin DN: A history of boron neutron capture therapy of
brain tumors. Brain 114: 1609–1629, 1991
2. Hawthorne MH: The role of chemistry in the development
of boron neutron capture therapy of cancer. Angew Chem
Int Ed Engl 32: 950–984, 1993
3. Chadwick J: The existence of a neutron. Proc Roy Soc Lon-
don 136: 692–708, 1932
4. Fermi E: Artificial radioactivity produced by neutron bom-
bardment. In: Holberg MA (ed) Les Prix Nobel in 1939.
Norstedt and Söner, Stockholm, 1939
5. Chadwick J, Goldhaber M: Disintegration by slow neutrons.
Nature 135: 65, 1935
6. Taylor JH, Goldhaber M: Detection of nuclear disintegra-
tion in a photographic emulsion. Nature 135: 34, 1935
7. Burcham WE, Goldhaber M: The disintegration of nitrogen
by slow neutrons. Proc Cambridge Philosophical Society 32:
632–636, 1936
8. Locher GL: Biological effects and therapeutic possibilities
of neutrons. Am J Roentgenol 36: 1–13, 1936
9. Kruger PG: Some biological effects of nuclear disintegra-
tion products on neoplastic tissues. Proc Nat Acad Sci 26:
181–192, 1940
10. Zahl PA, Cooper FS, Dunning JR: Some in vivo effects of
localized nuclear disintegration products on a transplanta-
ble mouse sarcoma. Proc Natl Acad Sci 26: 589–598, 1940
11. Zahl PA, Cooper FS: Physical and biological considerations
in the use of slow neutrons for cancer therapy. Radiology 37:
673–682, 1941
12. Sinex FM, Farr LE, Robertson JS: Dye carriers of boron for
neutron capture therapy. In: Quarterly Progress Report,
April 1–June 30. Brookhaven Library, Upton, NY, BNL 243
(S-18): 49, 1953
13. McClintock LA, Friedman MM: Utilization of antibody for
26
the localization of metals and dyes in the tissues. Prelimina-
ry report. Am J Roentgenol 54: 704–706, 1945
14. Knock FE: Perfusion of uranium-antibody complexes
forthe neutron capture therapy of tumors. Surg Gynec OB
109: 445–449, 1959
15. Tobias CA, Weumouth PP, Wasserman LR, Stapleton GE:
Some biological effects due to nuclear fission. Science 107:
115–118, 1948
16. Kanner MH, Barshall HH: Phys Rev 51: 372, 1940
17. Struxness EG, Luessenhop AJ, Bernard SR, Gallimore JC:
The distribution and excretion of hexavalent uranium in
man. Proc Int Conf Peaceful Uses of Atomic Energy 10: 186–
196, 1956
18. Luessenhop AJ, Gallimore JC, Sweet WH, Struxness EG,
Robinson J: The toxicity in man of hexavalent uranium fol-
lowing intravenous administration. Am J Roentgenol 79:
83–100, 1958
19. Selverstone B, Solomon AK, Sweet WH: Location of brain
tumors by means of radioactive phosphorus. JAMA 140:
277–278, 1949
20. Selverstone B, Sweet WH, Robinson CV: Clinical use of ra-
dioactive phosphorus in surgery of brain tumors. Ann Surg
130: 643–650, 1949
21. Sweet WH: The uses of nuclear disintegration in the diag-
nosis and treatment of brain tumor. New Engl J Med 245:
875–878, 1951
22. Sweet WH, Javid M: The possible use of slow neutrons plus
boron10 in therapy of intracranial tumors. Trans Am Neurol
Assoc 76: 60–63, 1951
23. Conger AD, Giles NH Jr: Cytogenetic effect of slow neu-
trons. Genetics 35: 397–419, 1950
24. Javid M, Brownell GL, Sweet WH: The possible use of neu-
tron-capturing isotopes such as boron10 in the treatment of
neoplasms. 2. Computation of the radiation energies and es-
timates of effects in normal and neoplastic brain. J Clin In-
vest 31: 604–610, 1952
25. Slatkin DN, McChesney DD, Wallace DW: A retrospective
study of 457 neurosurgical patients with cerebral malignant
glioma at the Massachusetts General Hospital, 1952–1981:
Implications for sequential trials of postoperative therapy.
In: Hatanaka H (ed) Proceeding Second International Sym-
posium Neutron Capture Therapy, Tokyo, October 18–20,
1985. Nishimura Co., Niigata, Japan 434–446, 1986
26. Farr LE, Sweet WH, Robertson JS, Foster CG, Locksley
HB, Sutherland DL, Mendelsohn ML, Stickley EE: Neu-
tron-capture therapy with boron in the treatment of glio-
blastoma multiforme. Am J Roentgenol 71: 279–293, 1954
27. Godwin JT, Farr LE, Sweet WH, Robertson JS: Pathological
study of 8 patients with glioblastoma multiforme treated by
neutron-capture therapy using boron10. Cancer 8: 601–615,
1955
28. Soloway AH: Correlation of drug penetration of brain and
chemical structure. Science 128: 1572–1574, 1958
Please indicate author’s corrections in blue, setting errors in red
121162 NEON ART.NO 734AS (590) ORD.NO 234590.Z
29. Soloway AH, Wright RL, Messer JR: Evaluation of boron-
compounds for use in neutron capture therapy of brain tu-
mors. I. Animal investigations. J Pharmacol Exp Ther 134:
117–122, 1961
30. Sweet WH, Soloway AH, Wright RL: Evaluation of boron
compounds for use in neutron capture therapy of brain tu-
mors. II. Studies in man. J Pharmacol Exp Ther 137: 263–
266, 1962
31. Asbury AK, Ojemann RG, Nielsen SL, Sweet WH: Neuro-
pathologic study of 14 cases of malignant brain tumor treat-
ed by boron-10 slow neutron capture radiation. J Neuro-
pathol Exper Neurol 31: 278–303, 1972
32. Kitao K: A method for calculating the absorbed dose near
interface from 10B(n,α) 7Li reaction. Radiat Res 61: 304–315,
1975
33. Rydin RA, Deutsch OL, Murray BW: The effect of geom-
etry on capillary wall dose for boron neutron capture ther-
apy. Phys Med Biol 21: 134–138, 1976
34. Slatkin DN, Stoner RD, Rosander KM, Kalef-Ezra JA, Lais-
sue JA: Central nervous system radiation syndrome in mice
from preferential 10B(n, α) 7Li irradition in brain vascula-
ture. Proc Natl Acad Sci 85: 4020–4024, 1988
35. Fairchild RG, Gabel D, Laster BH, Greenberg D, Kiszenick
W, Micca PL: Microanalytical techniques for boron analysis
using the 10B (n, α) 7Li reaction. Med Phys 13: 50–56, 1986
36. Soloway AH, Hatanaka H, Davis MA: Penetration of brain
and brain tumor. VII. Tumor-binding sulfhydryl boron com-
pounds. J Med Chem 10: 714–717, 1967
37. Tolpin EI, Wellum GR, Berley SA: Synthesis and chemistry
of mercaptoundecahydro-closo-dodecaborate (2−). Inorga-
nic Chem 17: 2867–2873, 1978
38. Hatanaka H: Clinical experience of boron neutron capture
therapy for gliomas – a comparison with conventional che-
mo-immuno-therapy. In: Hatanaka H (ed) Boron Neutron
Capture Therapy for Brain Tumors. Nishimura Co., Niigata,
Japan 349–379, 1986
39. Sano K: Preface. In: Hatanaka H (ed) Boron Neutron Cap-
ture Therapy for Brain Tumors. Nishimura Co., Niigata, Ja-
pan, 1986
40. Wilson CB: Glioblastoma: the past, the present and the fu-
ture. Clin Neurosurg 38: 32–48, 1992
41. Snyder HR, Reedy AJ, Lennarz WJ: Synthesis of aromatic
boronic acids. Aldehydo boronic acids and boronic acid ana-
log of tyrosine. J Am Chem Soc 80: 835, 1958
42. Mishima Y: Selective neutron capture therapy of cancer
cells using their specific metabolic activities – Melanoma as
prototype (Abstract). Sixth Symposium on Neutron Cap-
ture Therapy for Cancer. Kobe, Japan, October 31–Novem-
ber 4, 1994, p. 29
Address for offprints: W.H. Sweet, 309 Goddard Avenue Brook-
line, MA 02146, USA