1992, The British Journal of Radiology, 65, 66-71
Biological intercomparisons of neutron beams used for radiotherapy generated by p + ->Be in hospital-based cyclotrons By Eric J . Hall, DPhil, DSc, Myles Astor, PhD and David J . Brenner, PhD Center for Radiological Research, College of Physicians & Surgeons of Columbia University, New York, NY, USA (Received 7 May 1991 and in revised form 5 August 1991, accepted 4 September 1991) Keywords: Neutron therapy, Biological intercomparisons, Relative biological effectiveness
Abstract. The new generation of hospital-based neutron therapy facilities involve cyclotrons using protons on beryllium. The spectrum of neutrons produced includes a large and variable proportion of low-energy neutrons that are poorly penetrating but biologically effective. Cells cultured in vitro were used to compare the three US facilities at Seattle, M.D. Anderson and UCLA, together with the UK facility at Clatter bridge. Cyclotrons were compared within a given experiment on the same day using cells from a common suspension. Among the three US facilities, the relative potency factor at a depth of 25 mm differs by about 11%, with Seattle the least and UCLA the most biologically effective. Clatterbridge was compared directly with M.D. Anderson and found to be less effective by about 5%; it has a slightly lower biological effectiveness than any of the US facilities. There is evidence for an increased biological effectiveness in the build-up region, which reduces the effective skin sparing potential. There is not much difference in build-up between the three US facilities. Using the proton-on-beryllium neutron production process results in a wide spectrum of neutrons with a large but variable low-energy component. The biological effectiveness of the beam depends on target design and thickness as well as the design of the collimating system. Consequently the biological effectiveness of neutron beams generated by this process must be assessed on an individual basis. It cannot be assumed that because cyclotrons have similar accelerating energies that the relative biological effectiveness will be the same.
Stone and his colleagues at Berkeley were the first to use neutrons for radiotherapy in the late 1930s; this early study resulted in unacceptable late sequelae and was terminated abruptly by the entry of the United States (US) into World War II (Stone, 1948; Sheline et al, 1971). Interest in the use of neutrons for the treatment of human cancer was rekindled in the late 1950s at the Hammersmith Hospital in London, where patient treatment was not started until an impressive amount of careful experimentation had been completed in both radiological physics and radiation biology. This well documented work formed the basis for the first clinical trials (Fowler et al, 1963; Bewley, 1963; Catterall, 1974, 1977). The early success of the Hammersmith group inspired interest on both sides of the Atlantic as well as in Japan in the possibility of establishing additional clinical neutron facilities (Griffin et al, 1979; Tsunemoto et al, 1979). In the US, these took the form of large cyclotrons, initially built for high-energy physics research and therefore often located in physics departments remote from major hospitals and having a fixed horizontal or vertical beam (Griffin et al, 1979). Under these circumstances it proved difficult to treat effectively a large number of patients in randomized clinical trials. The new generation of neutron therapy machines, therefore, in both the US and Great Britain are hospital-based cyclotrons, having good depth-dose characteristics, a high output and an isocentric mount. A more compact machine that can be accommodated in a hospital is made possible by 66
changing the neutron production process from deuterons on beryllium to protons on beryllium (Bonnett et al, 1988). A cyclotron accelerating protons to a given energy is, of course, much smaller and more compact than one accelerating deuterons. However, other complications arise from the change of the accelerated particle. In the case of d + ->• Be, the neutrons produced have a spectrum with a single peak, while the p + -> Be process results in a neutron beam with a large component of both low-energy and high-energy neutrons. The beams are, therefore, very different. Some, though not all, of the beams in clinical use have been characterized spectroscopically using time of flight measurements (e.g. Crout et al, 1991). However, one problem with this technique is that it is limited to energies above ~ 2 MeV, thus the low-energy component is rarely known. Intercomparisons Because of the substantial cost and investment of effort involved in the implementation of neutron clinical trials, the value of full co-operation between the few centres using these particles has been recognized from the outset. A great deal of effort was expended by the physicists at the various installations used for neutron therapy to achieve compatible dosimetry, and intercomparisons indicate agreement in the measurement of physical dose to within a few per cent (Smith et al, 1975; Broerse et al, 1981). However, compatible dosimetry does not of itself allow the radiotherapist to compare dosage schedules The British Journal of Radiology, January 1992
Neutron inter comparisons
directly, since the varying neutron energies result in a different relative biological effectiveness (RBE) for each neutron beam. In order to facilitate the pooling of experience and a comparison of clinical results, a number of individuals have undertaken biological intercomparisons using portable model systems (Hall et al, 1975, 1979; Todd et al, 1975; Hall, 1977; Hall & Kellerer, 1979). Previous intercomparisons of the earlier generation of neutron facilities, which mostly employed the d + -»• Be production process, showed RBE to be a function largely of the energy of the accelerated deuteron. Based on experiments performed in the US, Great Britain and Japan, RBE values plotted as a function of deuteron energy fell around a common line with RBE decreasing with increasing energy (Hall & Kellerer, 1979; Hall et al, 1979). The same cannot be said for the new generation of hospital-based cyclotrons where neutrons are generated by the p + -»• Be reaction; RBE values from different facilities do not show a simple relationship to proton energy. It is true that RBE varies linearly with proton energy if experiments are performed on a single machine in which the energy is varied (Beauduin et al, 1989), but it is no longer true when the range of energies involves different facilities. The obvious conclusion from this observation is that factors other than proton energy are of more importance—these include target design and collimator configuration (Hall et al, 1982; Hall et al, 1983; Hornsey et al, 1988; Bewley et al, 1989). The present paper describes biological comparisons between three hospital-based neutron facilities in the US and one in Great Britain. The characteristics of the machines are listed in Table I. Materials and methods
Chinese hamster V79 cells cultured in vitro were used in this series of comparison experiments. Cells were grown using standard tissue-culture techniques in Ham's F10 supplemented with 10% fetal bovine serum (Sterile Systems), antibiotics (Gibco), and L-Glutamine (Gibco). In order to exploit fully the precision of which the in vitro system is capable, attached cells were used and cyclotrons were compared within a given experiment on the same day. For cells in culture, variations within an
experiment are much smaller than between experiments. Within a given experiment, the repeatability between replicate flasks is limited only by the counting statistics, and since large numbers of colonies can be used, the accuracy is of the order of a few per cent. By contrast, when experiments are repeated on separate occasions, it is not unusual for the cell surviving fraction at a given dose level to differ by a factor of 2. For this reason, the system was adopted of comparing facilities within a given experiment. The details have been published previously (Hall et al, 1982, 1983). For each experiment, exponentially growing V79 cells were harvested, spun down and counted, and plated into 25-cm2 (Corning) tissue-culture flasks. The cells were allowed to attach overnight. On the morning of the experiment, the flasks were completely filled with ice-cold medium (to stop the cells from cycling) and placed into specially constructed, temperaturecontrolled carriers for transport to the respective cyclotrons. The carriers consist of standard pressure cookers sealed to avoid changes of pressure when aircraft takeoff and land. The pressure cooker is surrounded by insulation and frozen material. The intent is to maintain the temperature within the range 10-18°C. It was found by trial and error that a temperature in this range preserved plating efficiency at 80-90% and prevented cells from moving through the cell cycle. If the temperature is allowed to exceed 22°C, cells begin to move through the cycle. One set of flasks was flown to each of the facilities to be compared. Experiment I In this experiment the three US facilities were compared, namely UCLA, Seattle and M. D. Anderson. The question that prompted the experiment was the clinical observation that skin reactions appeared to be more severe at some facilities than others. Flasks were irradiated simultaneously at the three facilities using the lucite phantom illustrated in Fig. 1. Cells were irradiated at depths of 1 mm (the thickness of the tissue culture flask) and at 25 mm. The doses were prescribed at a depth of 25 mm, which is close to the build-up depth, based on measurements with an ionization chamber in the phantom. Side view dosimetry
Table I. Characteristics of the four neutron facilities studied Energy of accelerated proton (MeV)
Location
Manufacturer
UCLA, Los Angeles, US
Cyclotron Corporation Cyclotron Corporation Scanditronix
45
Scanditronix
62
M.D. Anderson Hospital, Houston, US University of Washington, Seattle, US Clatterbridge, UK
Vol. 65, No. 769
Cell layer /
Thick block
42 50
Irradiation Facility for Cells at Depths of 1mm and 2.5 cm
Solid lucite .with chamber cut-out
Thick block Lucite Plate toAccommodate a 1cc EG & G Ionization Chamber
Figure 1. Set-up for irradiation of monolayers of cells at depths of 1 mm and 25 mm at the three US neutron facilities.
67
E. J. Hall, M. As tor and D. J. Brenner
Irradiation facility for cells at depths of 2 cm & 12 cm
Cavities in lucite phantom to accommodate a 1 cc EG & G ionization chamber
Figure 2. Set-up for irradiation of monolayers of cells at depths of 20 and 120 mm at neutron facilities in the US and United Kingdom. Experiment II In this experiment, the UK facility at Clatterbridge was compared with M.D. Anderson. This experiment addressed a somewhat different question, namely the possibility of a changing neutron RBE with depth. Flasks were irradiated at depths of 20 and 120 mm using the lucite phantom illustrated in Fig. 2. Irradiations were performed at Clatterbridge with and without the addition of a hydrogenous filter. The M. D. Anderson beam is always filtered. Following irradiation in either experiment, the flasks were returned to the carriers, transported back to Columbia and placed in an incubator. The exception was Clatterbridge, where the cells were incubated on site. Cells were incubated for a period of 6 days to allow
for colony formation, after which the clones were fixed and stained with formalin and giemsa, and scored. A colony was scored as a surviving cell if it contained 50 or more cells. The survival data were analysed using curves of the same shape, with a dose modifying factor which was varied from beam to beam to fit the data. A single dose factor is then a measure of the RBE or relative potency, between the beams, based on the data accumulated over the entire dose range. It is only a meaningful concept when comparing two neutron beams that are close in energy. When comparing, for example, neutrons with X-rays, RBE clearly varies with dose. To be specific our technique for comparing one set of survival data with another (or others) involves fitting the survival data to the expressions SID) = exp(-a/;Z)-^ 2 Z) 2 ) where the subscript i refers to different irradiation protocols (e.g. different depths or different machines). Thus £ is an estimator of the dose modifying effect, of one radiation relative to another. It is sometimes referred to as the "relative potency" (Hall et al, 1982). The estimation of the parameters [a, /?, f] was achieved using the maximum likelihood technique described elsewhere (Brenner & Hall, 1991). Results Experiment I The survival curves for cells irradiated at a depth of 25 mm at UCLA, Seattle and M. D. Anderson are shown in Fig. 3a. The relative potency factors between the three facilities at a depth of 25 mm, calculated as described above, are shown in Table II. It is clear from Table II and Fig. 3a that the UCLA machine is biologi43
1-
lq
0)
o
6
• l - l
-p O
^
2 o.N
0.1-
00 •
i-H
>
•>
o.oH
0.01-
Seattle
Seattle
Texas
Texas
25 mm depth o
0.0014
Dose (Gy)
6
0.001-
0
1 mm depth
UCLA 2
4
6
8
Dose delivered at 25 mm depth (Gy) (b)
(a) Figure 3. (a) Survival data for V79 Chinese hamster cells exposed to graded doses of neutrons at a depth of 25 mm at the three US neutron facilities. In this and subsequent figures the curves are fits to the linear-quadratic equation as described in the text. In addition, in all figures the error bars are standard errors determined by binomial statistics, (b) Survival data for V79 Chinese hamster cells exposed at a depth of 1 mm at the three US neutron facilities. Doses were prescribed at a depth of 25 mm and are thus roughly 25-30% larger than those delivered at 1 mm (see text). 68
The British Journal of Radiology, January 1992
Neutron inter comparisons Table II. Relative potency factors for the three US facilities at a depth of 25 mm 1.0 1.030 + 0.047 1.115 + 0.039
Seattle M.D. Anderson UCLA
Table "III. Biological build-up factors, i.e. ratio of biologically effective doses relative at 1 mm to 25 mm (see text) M.D. Anderson Seattle UCLA
0.94 + 0.03 0.97 + 0.03 0.92 + 0.04
cally more effective per unit dose than the two other US facilities.
The survival curves for cells irradiated at a depth of 1 mm at the three facilities are shown in Fig. 3b; the doses shown are those prescribed at a depth of 25 mm. Table III shows the biological build-up factors for the three facilities, i.e. the ratio of biologically effective doses at 1 mm to 25 mm. These ratios are, in effect, the product of two factors, namely the dose build-up factor and the relative potencies of the radiation between the depths of 1 and 25 mm. The physical dose build-up factor from 1 mm for a large field of 42 or 66 MeV p + -> Be neutrons has been reported to be in the range 0.75-0.8 (Mijnheer et al, 1978; Awschalom & Rosenberg, 1981; Horton et al, 1988). Since, the biological build-up factors are in the range from 0.92 to 0.96, this implies a relative potency of perhaps 1.2-1.3 at 1 mm compared with 25 mm. Such an increase probably reflects the soft neutron component of the beam, and will vary strongly with individual irradiation conditions. Experiment II Cell survival curves for cells irradiated at a depth of 20 mm at M.D. Anderson and Clatterbridge are shown
in Figs 4 and 5. In both cases the M. D. Anderson beam was filtered by a hydrogenous filter; all data in Fig. 4 relate to a filtered beam at Clatterbridge, while the data in Fig. 5 relate to an unfiltered beam at Clatterbridge. Figure 6 contains data from Clatterbridge for irradiations at depths of 20 and 120 mm. Relative potency factors from the data shown in Figs 4, 5 and 6 are summarized in Table IV. There are several conclusions to be drawn. First, the M. D. Anderson beam is biologically more effective than the Clatterbridge beam by about 4-7%. Second, the addition of a hydrogenous filter makes little difference to the Clatterbridge beam, nor does RBE vary much with depth. This confirms earlier reports concerning Clatterbridge, but is quite different to the situation at other facilities using the p + -• Be reaction (Bewley et al, 1989; Hall et al, 1982, 1983).
0.01; Clatterbridge (unfiltered) Texas (filtered)
CO
0.001 2
4
6
Dose (Gy) Figure 5. Survival data for V79 Chinese hamster cells irradiated with graded doses of neutrons at a depth of 20 mm with an unfiltered beam at Clatterbridge and a filtered beam at M. D. Anderson Hospital, Houston, Texas.
0.1:
01-
I •
\
0.01^
tio
\
Filtered beams 20 m m depth
~_ '_ x - A 0.001-
i
i
1
3 CO
1
Dose (Gy) Figure 4. Survival data for V79 Chinese hamster cells irradiated with graded doses of neutrons at a depth of 20 mm with filtered beams at Clatterbridge and M. D. Anderson Hospital, Houston. Vol. 65, No. 769
Clatterbridge
0.01-
V
C latterbr idge Texas
x
20 m m
A
120 m m
0.001 4
6
Dose (Gy) Figure 6. Survival data for V79 Chinese hamster cells irradiated with graded doses of neutrons at depths of 20 and 120 mm at Clatterbridge, UK.
69
E. J. Hall, M. Astor and D. J. Brenner Table IV. Relative potency factors for the Clatterbridge and M.D. Anderson neutron therapy facilities 20 mm depth Clatterbridge (filtered) Clatterbridge unfiltered
1.0 1.029 + 0.038
Clatterbridge filtered M.D. Anderson filtered
1.0 1.069 ±0.033
Clatterbridge unfiltered M.D. Anderson filtered
1.0 1.044 + 0.026
Clatterbridge Clatterbridge
20 mm: 1.0 120 mm: 1.007 ±0.019
Discussion
The general conclusion that can be drawn from the data in this paper, together with previous experience, is that RBE values of cyclotron-produced neutron beams using the p + -> Be reaction need to be assessed on an individual basis. It cannot be assumed that because different cyclotrons have the same accelerating energy, the RBE will be the same. Nor can it it be assumed that RBE will necessarily decrease with increasing energy of the accelerated proton. It further follows that if a common protocol is used at several facilities under the auspices of a co-operative group, it is not necessarily appropriate to use a common dose level, however carefully the physical dosimetry is compared. It has previously been reported that for high-energy neutron beams generated by p + ->• Be, there may be a change of RBE with depth, which can be reduced or eliminated by the use of a hydrogenous filter, and that this change of RBE is highly variable between different facilities (Hall et al, 1983; Hornsey et al, 1988). The data in the present paper may be briefly summarized as follows: amongst the three US neutron facilities, relative potencies at a depth of 25 mm differ by about 11 % with Seattle the least biologically effective and UCLA the most effective. Clatterbridge is less effective than M. D. Anderson by about 5% and is therefore slightly less effective than any of the US machines. In terms of the biological build-up between 1 mm and 25 mm, i.e. the ratio of biological effective doses between these depths, there is a small variation between the three US machines with UCLA showing the largest difference between depths and M. D. Anderson the smallest. Based on an assumed dose build-up factor of 0.75-0.8, the relative potency for the neutrons at 1 mm is probably about 1.2-1.3 compared with at 25 mm (Table III). The changing RBE with depth and the finding in the present paper that RBE is not a simple function of proton energy are both manifestations of the shape of the neutron spectum generated by the p + -»• Be process. A wide spectrum of neutrons is produced including a substantial (but variable) low-energy component, which is highly effective biologically and is filtered out preferentially in the first few centimetres of absorbing material. The RBE of the neutron beam, particularly at 70
shallow depths, is critically dependent on target thickness and design, as well as the design of the collimating system, all factors that have major influence on the magnitude of the low-energy neutron component of the beam. As discussed above, this low-energy component is rarely measured, as it is not accessible by time-of-flight techniques. It is, however, measurable using activation techniques, as demonstrated by Greenwood et al (1979) for a 60 MeV d+ -»• Be beam. Spectral measurements and the determination of microdosimetric quantities in a few of the high-energy beams confirm the explanation that the RBE fluctuations are a consequence of a variable low-energy neutron component, but insufficient physical measurements have been performed (certainly at the US facilities) to allow a detailed correlation to be made between physical and biological parameters. Acknowledgments Intercomparison experiments of this kind are only possible with the unstinting help and co-operation of colleagues at all of the institutions involved. These include: Drs David Bewley, John Parnall and David Bonnett at Clatterbridge; Drs Robert Parker, James Smathers and Ms Kathy Mason at UCLA; Drs Lester Peters, John Horton and Pat Stafford at M. D. Anderson Hospital, Houston, Texas; Drs Tom Griffin, Juri Eenmaa and Peter Wootton at the University of Washington, Seattle. This investigation was supported by Grant Numbers CA 12536 and CA 18506 from the National Cancer Institute. References AWSCHALOM, M. & ROSENBERG, I., 1981. Characterization of a
p (66) Be (49) neutron therapy beam II. Skin-sparing and dose transition effects. Medical Physics, 8, 105-107. BEAUDUIN, M., GUEULETTE, J., VYNCKIER, S. & WAMBERSIE, A.,
1989. Radiobiological intercomparison of clinical neutron beams for growth inhibition in Vicia faba bean roots. Radiation Research, 117, 245-250. BEWLEY, D. K., 1963. Pre-therapeutic experiments with the fast neutron beam from the Medical Research Council cyclotron. II. Physical aspects of the fast neutron beam. British Journal of Radiology, 36, 81-88. BEWLEY, D. K., CULLEN, B. M., ASTOR, M. HALL, E. J., BLAKE, S. W., BONNETT, D. E. & ZAIDER, M., 1989. Changes in
biological effectiveness of the neutron beam at Clatterbridge (62 MeV p on Be) measured with cells in vitro. British Journal of Radiology, 62, 344-347. BONNETT, D. E., BLAKE, S. W., SHAW, J. E., & BEWLEY, D. K.,
1988. The Clatterbridge high-energy neutron therapy facility: specification and performance. British Journal of Radiology, 62, 38-46. BRENNER, D. J. & HALL, E. J., 1991. Conditions for the
equivalence of continuous to pulsed low dose-rate brachytherapy. International Journal of Radiation Oncology, Biology, Physics, 20, 181-190. BROERSE, J. J., MIJNHEER, B. J. & WILLIAM, J. R., 1981.
European protocol for neutron dosimetry for external beam therapy. British Journal of Radiology, 54, 882-898. CATTERALL, M., 1974. A report on three years' fast neutron therapy from the Medical Research Council's cyclotron at Hammersmith Hospital, London. Cancer, 34, 91-95. CATTERALL, M., 1977. The results of randomized and other clinical trials of fast neutrons from the Medical Research Council Cyclotron, London. International Journal of Radiation Oncology, Biology, Physics, 3, 247-253. The British Journal of Radiology, January 1992
Neutron inter comparisons CROUT, N. M. J., FLETCHER, J. G., GREEN, S., SCOTT, M. C. &
TAYLOR, G. C , 1991. In situ neutron spectrometry to 60 MeV in a water phantom exposed to a cancer therapy beam. Physics in Medicine and Biology, 36, 507-519.
HALL, E. J., ZAIDER, M., BIRD, R. & ROBERTS, W., 1983.
Radiobiological studies with therapeutic neutron beams generated by p + - > B e or d + - » B e . British Journal of Radiology, 56, 349-350.
FOWLER, J. F., MORGAN, R. L. & WOOD, C. A. P., 1963.
HORNSEY, S., MYERS, R., PARNELL, C. J., BONNETT, D. E.,
Pre-therapeutic experiments with the fast neutron beam from the Medical Research Council cyclotron. I. The biological and physical advantages and problems of neutron therapy. British Journal of Radiology, 36, 77-80.
biological effectiveness with depth of the Clatterbridge neutron beam. British Journal of Radiology, 61, 1058-1062.
GREENWOOD, L. R., HEINRICH, R. R., SALTMARSH, M. J. &
FULMER, C. B., 1979. Integral tests of neutron activation cross sections in a 9Be(d,n) field at Ed = 40 MeV. Nuclear Science and Engineering, 72, 175-190. GRIFFIN, T., BLASKO, J. & LARAMORE, G., 1979. Results of fast
neutron beam radiotherapy pilot studies at the University of Washington. In High LET Radiations in Clinical Radiotherapy (Supplement to European Journal of Cancer), 23-30. HALL, E. J., 1977. Radiobiological intercomparison in vivo and in vitro. International Journal of Radiation Oncology, Biology, Physics 3, 195-201. HALL, E. J. & KELLERER, A., 1979. Review of RBE data for
cells in culture. In High LET Radiations in Clinical Radiotherapy (Supplement to European Journal of Cancer), 171-174. HALL, E. J., ROIZIN-TOWLE, L., THEUS, R. B. & AUGUST, L. S.,
1975. Radiobiological properties of high energy cyclotron produced neutrons used for radiotherapy. Radiology, 117, 173-178. HALL, E. J., WITHERS, H. R., GERACI, J. P., MEYN, R. E., RASEY, J., TODD, P. & SHELINE, G. E., 1979. Radiobiological
intercomparisons of fast neutron beams used for therapy in Japan and the United States. International Journal of Radiation Oncology, Biology, Physics, 5, 227-233. HALL, E. J., ZAIDER, M., BIRD, R. & ASTOR, M., 1982.
Radiobiological studies with therapeutic neutron beams generated by p + -> Be or d + - » B e . British Journal of Radiology, 55, 640-644.
Vol. 65, No. 769
BLAKE, S. W. & BEWLEY, D. K., 1988. Changes in relative
HORTON J. L., OTTE, V. A., SCHULTEISS, T. E., STAFFORD,
P. M., SUN, T. & ZERMENO, A., 1988. Physical characteristics of the M. D. Anderson Hospital clinical neutron beam. Radiotherapy and Oncology, 13, 17-22. MIJNHEER, B. J., ZOETLIEF, J. & BROERSE, J. J., 1978. Build-up
and depth dose characteristics of different fast neutron beams relevant to radiotherapy. British Journal of Radiology, 51, 122-126. SHELINE, G. E., PHILLIPS, T. L., FIELD, S. B., BRENNAN, J. T. &
RAVENTOS, A., 1971. Effects of fast neutrons on human skin. American Journal of Roentgenology, 111, 31-41. SMITH, A. R., ALMOND, P. R., SMATHERS, J. B., OTTE, V. A., ATTIX, F. H., THEUS, R. B., WOOTTON, P., BICHSEL, H., EENMAA, J., WILLIAMS, D., BEWLEY, D. K. & PARNELL, C. J.,
1975. Dosimetry intercomparisons between fast-neutron radiotherapy facilities. Medical Physics, 2, 195-200. STONE, R. S., 1948. Neutron therapy and specific ionizations. American Journal of Roentgenology, 59, 771-779. TODD, P., GERACI, J. P., FURCINITTI, P. S., Rossi, R. M., MIKAGE, F., THEUS, P. & SCHROY, C. B., 1975. Comparison
of the effects of various cyclotron-produced fast neutrons on the reproductive capacity of cultured human kidney (T-l) cells. International Journal of Radiation Oncology, Biology, Physics, 4, 1015-1022. TSUNEMOTO, H., UMEGAKI, Y., KUTSUTANI, Y., ARAI, T., MORITA, S., KURISU, A., KAWASHIMA, K. & MARUYAMA, T.,
1979. Results of clinical applications of fast neutrons in Japan. In High LET Radiations in Clinical Radiotherapy (Supplement to European Journal of Cancer), 75-78.
71
Biological intercomparisons of neutron beams used for radiotherapy generated by p + ->Be in hospital-based cyclotrons By Eric J . Hall, DPhil, DSc, Myles Astor, PhD and David J . Brenner, PhD Center for Radiological Research, College of Physicians & Surgeons of Columbia University, New York, NY, USA (Received 7 May 1991 and in revised form 5 August 1991, accepted 4 September 1991) Keywords: Neutron therapy, Biological intercomparisons, Relative biological effectiveness
Abstract. The new generation of hospital-based neutron therapy facilities involve cyclotrons using protons on beryllium. The spectrum of neutrons produced includes a large and variable proportion of low-energy neutrons that are poorly penetrating but biologically effective. Cells cultured in vitro were used to compare the three US facilities at Seattle, M.D. Anderson and UCLA, together with the UK facility at Clatter bridge. Cyclotrons were compared within a given experiment on the same day using cells from a common suspension. Among the three US facilities, the relative potency factor at a depth of 25 mm differs by about 11%, with Seattle the least and UCLA the most biologically effective. Clatterbridge was compared directly with M.D. Anderson and found to be less effective by about 5%; it has a slightly lower biological effectiveness than any of the US facilities. There is evidence for an increased biological effectiveness in the build-up region, which reduces the effective skin sparing potential. There is not much difference in build-up between the three US facilities. Using the proton-on-beryllium neutron production process results in a wide spectrum of neutrons with a large but variable low-energy component. The biological effectiveness of the beam depends on target design and thickness as well as the design of the collimating system. Consequently the biological effectiveness of neutron beams generated by this process must be assessed on an individual basis. It cannot be assumed that because cyclotrons have similar accelerating energies that the relative biological effectiveness will be the same.
Stone and his colleagues at Berkeley were the first to use neutrons for radiotherapy in the late 1930s; this early study resulted in unacceptable late sequelae and was terminated abruptly by the entry of the United States (US) into World War II (Stone, 1948; Sheline et al, 1971). Interest in the use of neutrons for the treatment of human cancer was rekindled in the late 1950s at the Hammersmith Hospital in London, where patient treatment was not started until an impressive amount of careful experimentation had been completed in both radiological physics and radiation biology. This well documented work formed the basis for the first clinical trials (Fowler et al, 1963; Bewley, 1963; Catterall, 1974, 1977). The early success of the Hammersmith group inspired interest on both sides of the Atlantic as well as in Japan in the possibility of establishing additional clinical neutron facilities (Griffin et al, 1979; Tsunemoto et al, 1979). In the US, these took the form of large cyclotrons, initially built for high-energy physics research and therefore often located in physics departments remote from major hospitals and having a fixed horizontal or vertical beam (Griffin et al, 1979). Under these circumstances it proved difficult to treat effectively a large number of patients in randomized clinical trials. The new generation of neutron therapy machines, therefore, in both the US and Great Britain are hospital-based cyclotrons, having good depth-dose characteristics, a high output and an isocentric mount. A more compact machine that can be accommodated in a hospital is made possible by 66
changing the neutron production process from deuterons on beryllium to protons on beryllium (Bonnett et al, 1988). A cyclotron accelerating protons to a given energy is, of course, much smaller and more compact than one accelerating deuterons. However, other complications arise from the change of the accelerated particle. In the case of d + ->• Be, the neutrons produced have a spectrum with a single peak, while the p + -> Be process results in a neutron beam with a large component of both low-energy and high-energy neutrons. The beams are, therefore, very different. Some, though not all, of the beams in clinical use have been characterized spectroscopically using time of flight measurements (e.g. Crout et al, 1991). However, one problem with this technique is that it is limited to energies above ~ 2 MeV, thus the low-energy component is rarely known. Intercomparisons Because of the substantial cost and investment of effort involved in the implementation of neutron clinical trials, the value of full co-operation between the few centres using these particles has been recognized from the outset. A great deal of effort was expended by the physicists at the various installations used for neutron therapy to achieve compatible dosimetry, and intercomparisons indicate agreement in the measurement of physical dose to within a few per cent (Smith et al, 1975; Broerse et al, 1981). However, compatible dosimetry does not of itself allow the radiotherapist to compare dosage schedules The British Journal of Radiology, January 1992
Neutron inter comparisons
directly, since the varying neutron energies result in a different relative biological effectiveness (RBE) for each neutron beam. In order to facilitate the pooling of experience and a comparison of clinical results, a number of individuals have undertaken biological intercomparisons using portable model systems (Hall et al, 1975, 1979; Todd et al, 1975; Hall, 1977; Hall & Kellerer, 1979). Previous intercomparisons of the earlier generation of neutron facilities, which mostly employed the d + -»• Be production process, showed RBE to be a function largely of the energy of the accelerated deuteron. Based on experiments performed in the US, Great Britain and Japan, RBE values plotted as a function of deuteron energy fell around a common line with RBE decreasing with increasing energy (Hall & Kellerer, 1979; Hall et al, 1979). The same cannot be said for the new generation of hospital-based cyclotrons where neutrons are generated by the p + -»• Be reaction; RBE values from different facilities do not show a simple relationship to proton energy. It is true that RBE varies linearly with proton energy if experiments are performed on a single machine in which the energy is varied (Beauduin et al, 1989), but it is no longer true when the range of energies involves different facilities. The obvious conclusion from this observation is that factors other than proton energy are of more importance—these include target design and collimator configuration (Hall et al, 1982; Hall et al, 1983; Hornsey et al, 1988; Bewley et al, 1989). The present paper describes biological comparisons between three hospital-based neutron facilities in the US and one in Great Britain. The characteristics of the machines are listed in Table I. Materials and methods
Chinese hamster V79 cells cultured in vitro were used in this series of comparison experiments. Cells were grown using standard tissue-culture techniques in Ham's F10 supplemented with 10% fetal bovine serum (Sterile Systems), antibiotics (Gibco), and L-Glutamine (Gibco). In order to exploit fully the precision of which the in vitro system is capable, attached cells were used and cyclotrons were compared within a given experiment on the same day. For cells in culture, variations within an
experiment are much smaller than between experiments. Within a given experiment, the repeatability between replicate flasks is limited only by the counting statistics, and since large numbers of colonies can be used, the accuracy is of the order of a few per cent. By contrast, when experiments are repeated on separate occasions, it is not unusual for the cell surviving fraction at a given dose level to differ by a factor of 2. For this reason, the system was adopted of comparing facilities within a given experiment. The details have been published previously (Hall et al, 1982, 1983). For each experiment, exponentially growing V79 cells were harvested, spun down and counted, and plated into 25-cm2 (Corning) tissue-culture flasks. The cells were allowed to attach overnight. On the morning of the experiment, the flasks were completely filled with ice-cold medium (to stop the cells from cycling) and placed into specially constructed, temperaturecontrolled carriers for transport to the respective cyclotrons. The carriers consist of standard pressure cookers sealed to avoid changes of pressure when aircraft takeoff and land. The pressure cooker is surrounded by insulation and frozen material. The intent is to maintain the temperature within the range 10-18°C. It was found by trial and error that a temperature in this range preserved plating efficiency at 80-90% and prevented cells from moving through the cell cycle. If the temperature is allowed to exceed 22°C, cells begin to move through the cycle. One set of flasks was flown to each of the facilities to be compared. Experiment I In this experiment the three US facilities were compared, namely UCLA, Seattle and M. D. Anderson. The question that prompted the experiment was the clinical observation that skin reactions appeared to be more severe at some facilities than others. Flasks were irradiated simultaneously at the three facilities using the lucite phantom illustrated in Fig. 1. Cells were irradiated at depths of 1 mm (the thickness of the tissue culture flask) and at 25 mm. The doses were prescribed at a depth of 25 mm, which is close to the build-up depth, based on measurements with an ionization chamber in the phantom. Side view dosimetry
Table I. Characteristics of the four neutron facilities studied Energy of accelerated proton (MeV)
Location
Manufacturer
UCLA, Los Angeles, US
Cyclotron Corporation Cyclotron Corporation Scanditronix
45
Scanditronix
62
M.D. Anderson Hospital, Houston, US University of Washington, Seattle, US Clatterbridge, UK
Vol. 65, No. 769
Cell layer /
Thick block
42 50
Irradiation Facility for Cells at Depths of 1mm and 2.5 cm
Solid lucite .with chamber cut-out
Thick block Lucite Plate toAccommodate a 1cc EG & G Ionization Chamber
Figure 1. Set-up for irradiation of monolayers of cells at depths of 1 mm and 25 mm at the three US neutron facilities.
67
E. J. Hall, M. As tor and D. J. Brenner
Irradiation facility for cells at depths of 2 cm & 12 cm
Cavities in lucite phantom to accommodate a 1 cc EG & G ionization chamber
Figure 2. Set-up for irradiation of monolayers of cells at depths of 20 and 120 mm at neutron facilities in the US and United Kingdom. Experiment II In this experiment, the UK facility at Clatterbridge was compared with M.D. Anderson. This experiment addressed a somewhat different question, namely the possibility of a changing neutron RBE with depth. Flasks were irradiated at depths of 20 and 120 mm using the lucite phantom illustrated in Fig. 2. Irradiations were performed at Clatterbridge with and without the addition of a hydrogenous filter. The M. D. Anderson beam is always filtered. Following irradiation in either experiment, the flasks were returned to the carriers, transported back to Columbia and placed in an incubator. The exception was Clatterbridge, where the cells were incubated on site. Cells were incubated for a period of 6 days to allow
for colony formation, after which the clones were fixed and stained with formalin and giemsa, and scored. A colony was scored as a surviving cell if it contained 50 or more cells. The survival data were analysed using curves of the same shape, with a dose modifying factor which was varied from beam to beam to fit the data. A single dose factor is then a measure of the RBE or relative potency, between the beams, based on the data accumulated over the entire dose range. It is only a meaningful concept when comparing two neutron beams that are close in energy. When comparing, for example, neutrons with X-rays, RBE clearly varies with dose. To be specific our technique for comparing one set of survival data with another (or others) involves fitting the survival data to the expressions SID) = exp(-a/;Z)-^ 2 Z) 2 ) where the subscript i refers to different irradiation protocols (e.g. different depths or different machines). Thus £ is an estimator of the dose modifying effect, of one radiation relative to another. It is sometimes referred to as the "relative potency" (Hall et al, 1982). The estimation of the parameters [a, /?, f] was achieved using the maximum likelihood technique described elsewhere (Brenner & Hall, 1991). Results Experiment I The survival curves for cells irradiated at a depth of 25 mm at UCLA, Seattle and M. D. Anderson are shown in Fig. 3a. The relative potency factors between the three facilities at a depth of 25 mm, calculated as described above, are shown in Table II. It is clear from Table II and Fig. 3a that the UCLA machine is biologi43
1-
lq
0)
o
6
• l - l
-p O
^
2 o.N
0.1-
00 •
i-H
>
•>
o.oH
0.01-
Seattle
Seattle
Texas
Texas
25 mm depth o
0.0014
Dose (Gy)
6
0.001-
0
1 mm depth
UCLA 2
4
6
8
Dose delivered at 25 mm depth (Gy) (b)
(a) Figure 3. (a) Survival data for V79 Chinese hamster cells exposed to graded doses of neutrons at a depth of 25 mm at the three US neutron facilities. In this and subsequent figures the curves are fits to the linear-quadratic equation as described in the text. In addition, in all figures the error bars are standard errors determined by binomial statistics, (b) Survival data for V79 Chinese hamster cells exposed at a depth of 1 mm at the three US neutron facilities. Doses were prescribed at a depth of 25 mm and are thus roughly 25-30% larger than those delivered at 1 mm (see text). 68
The British Journal of Radiology, January 1992
Neutron inter comparisons Table II. Relative potency factors for the three US facilities at a depth of 25 mm 1.0 1.030 + 0.047 1.115 + 0.039
Seattle M.D. Anderson UCLA
Table "III. Biological build-up factors, i.e. ratio of biologically effective doses relative at 1 mm to 25 mm (see text) M.D. Anderson Seattle UCLA
0.94 + 0.03 0.97 + 0.03 0.92 + 0.04
cally more effective per unit dose than the two other US facilities.
The survival curves for cells irradiated at a depth of 1 mm at the three facilities are shown in Fig. 3b; the doses shown are those prescribed at a depth of 25 mm. Table III shows the biological build-up factors for the three facilities, i.e. the ratio of biologically effective doses at 1 mm to 25 mm. These ratios are, in effect, the product of two factors, namely the dose build-up factor and the relative potencies of the radiation between the depths of 1 and 25 mm. The physical dose build-up factor from 1 mm for a large field of 42 or 66 MeV p + -> Be neutrons has been reported to be in the range 0.75-0.8 (Mijnheer et al, 1978; Awschalom & Rosenberg, 1981; Horton et al, 1988). Since, the biological build-up factors are in the range from 0.92 to 0.96, this implies a relative potency of perhaps 1.2-1.3 at 1 mm compared with 25 mm. Such an increase probably reflects the soft neutron component of the beam, and will vary strongly with individual irradiation conditions. Experiment II Cell survival curves for cells irradiated at a depth of 20 mm at M.D. Anderson and Clatterbridge are shown
in Figs 4 and 5. In both cases the M. D. Anderson beam was filtered by a hydrogenous filter; all data in Fig. 4 relate to a filtered beam at Clatterbridge, while the data in Fig. 5 relate to an unfiltered beam at Clatterbridge. Figure 6 contains data from Clatterbridge for irradiations at depths of 20 and 120 mm. Relative potency factors from the data shown in Figs 4, 5 and 6 are summarized in Table IV. There are several conclusions to be drawn. First, the M. D. Anderson beam is biologically more effective than the Clatterbridge beam by about 4-7%. Second, the addition of a hydrogenous filter makes little difference to the Clatterbridge beam, nor does RBE vary much with depth. This confirms earlier reports concerning Clatterbridge, but is quite different to the situation at other facilities using the p + -• Be reaction (Bewley et al, 1989; Hall et al, 1982, 1983).
0.01; Clatterbridge (unfiltered) Texas (filtered)
CO
0.001 2
4
6
Dose (Gy) Figure 5. Survival data for V79 Chinese hamster cells irradiated with graded doses of neutrons at a depth of 20 mm with an unfiltered beam at Clatterbridge and a filtered beam at M. D. Anderson Hospital, Houston, Texas.
0.1:
01-
I •
\
0.01^
tio
\
Filtered beams 20 m m depth
~_ '_ x - A 0.001-
i
i
1
3 CO
1
Dose (Gy) Figure 4. Survival data for V79 Chinese hamster cells irradiated with graded doses of neutrons at a depth of 20 mm with filtered beams at Clatterbridge and M. D. Anderson Hospital, Houston. Vol. 65, No. 769
Clatterbridge
0.01-
V
C latterbr idge Texas
x
20 m m
A
120 m m
0.001 4
6
Dose (Gy) Figure 6. Survival data for V79 Chinese hamster cells irradiated with graded doses of neutrons at depths of 20 and 120 mm at Clatterbridge, UK.
69
E. J. Hall, M. Astor and D. J. Brenner Table IV. Relative potency factors for the Clatterbridge and M.D. Anderson neutron therapy facilities 20 mm depth Clatterbridge (filtered) Clatterbridge unfiltered
1.0 1.029 + 0.038
Clatterbridge filtered M.D. Anderson filtered
1.0 1.069 ±0.033
Clatterbridge unfiltered M.D. Anderson filtered
1.0 1.044 + 0.026
Clatterbridge Clatterbridge
20 mm: 1.0 120 mm: 1.007 ±0.019
Discussion
The general conclusion that can be drawn from the data in this paper, together with previous experience, is that RBE values of cyclotron-produced neutron beams using the p + -> Be reaction need to be assessed on an individual basis. It cannot be assumed that because different cyclotrons have the same accelerating energy, the RBE will be the same. Nor can it it be assumed that RBE will necessarily decrease with increasing energy of the accelerated proton. It further follows that if a common protocol is used at several facilities under the auspices of a co-operative group, it is not necessarily appropriate to use a common dose level, however carefully the physical dosimetry is compared. It has previously been reported that for high-energy neutron beams generated by p + ->• Be, there may be a change of RBE with depth, which can be reduced or eliminated by the use of a hydrogenous filter, and that this change of RBE is highly variable between different facilities (Hall et al, 1983; Hornsey et al, 1988). The data in the present paper may be briefly summarized as follows: amongst the three US neutron facilities, relative potencies at a depth of 25 mm differ by about 11 % with Seattle the least biologically effective and UCLA the most effective. Clatterbridge is less effective than M. D. Anderson by about 5% and is therefore slightly less effective than any of the US machines. In terms of the biological build-up between 1 mm and 25 mm, i.e. the ratio of biological effective doses between these depths, there is a small variation between the three US machines with UCLA showing the largest difference between depths and M. D. Anderson the smallest. Based on an assumed dose build-up factor of 0.75-0.8, the relative potency for the neutrons at 1 mm is probably about 1.2-1.3 compared with at 25 mm (Table III). The changing RBE with depth and the finding in the present paper that RBE is not a simple function of proton energy are both manifestations of the shape of the neutron spectum generated by the p + -»• Be process. A wide spectrum of neutrons is produced including a substantial (but variable) low-energy component, which is highly effective biologically and is filtered out preferentially in the first few centimetres of absorbing material. The RBE of the neutron beam, particularly at 70
shallow depths, is critically dependent on target thickness and design, as well as the design of the collimating system, all factors that have major influence on the magnitude of the low-energy neutron component of the beam. As discussed above, this low-energy component is rarely measured, as it is not accessible by time-of-flight techniques. It is, however, measurable using activation techniques, as demonstrated by Greenwood et al (1979) for a 60 MeV d+ -»• Be beam. Spectral measurements and the determination of microdosimetric quantities in a few of the high-energy beams confirm the explanation that the RBE fluctuations are a consequence of a variable low-energy neutron component, but insufficient physical measurements have been performed (certainly at the US facilities) to allow a detailed correlation to be made between physical and biological parameters. Acknowledgments Intercomparison experiments of this kind are only possible with the unstinting help and co-operation of colleagues at all of the institutions involved. These include: Drs David Bewley, John Parnall and David Bonnett at Clatterbridge; Drs Robert Parker, James Smathers and Ms Kathy Mason at UCLA; Drs Lester Peters, John Horton and Pat Stafford at M. D. Anderson Hospital, Houston, Texas; Drs Tom Griffin, Juri Eenmaa and Peter Wootton at the University of Washington, Seattle. This investigation was supported by Grant Numbers CA 12536 and CA 18506 from the National Cancer Institute. References AWSCHALOM, M. & ROSENBERG, I., 1981. Characterization of a
p (66) Be (49) neutron therapy beam II. Skin-sparing and dose transition effects. Medical Physics, 8, 105-107. BEAUDUIN, M., GUEULETTE, J., VYNCKIER, S. & WAMBERSIE, A.,
1989. Radiobiological intercomparison of clinical neutron beams for growth inhibition in Vicia faba bean roots. Radiation Research, 117, 245-250. BEWLEY, D. K., 1963. Pre-therapeutic experiments with the fast neutron beam from the Medical Research Council cyclotron. II. Physical aspects of the fast neutron beam. British Journal of Radiology, 36, 81-88. BEWLEY, D. K., CULLEN, B. M., ASTOR, M. HALL, E. J., BLAKE, S. W., BONNETT, D. E. & ZAIDER, M., 1989. Changes in
biological effectiveness of the neutron beam at Clatterbridge (62 MeV p on Be) measured with cells in vitro. British Journal of Radiology, 62, 344-347. BONNETT, D. E., BLAKE, S. W., SHAW, J. E., & BEWLEY, D. K.,
1988. The Clatterbridge high-energy neutron therapy facility: specification and performance. British Journal of Radiology, 62, 38-46. BRENNER, D. J. & HALL, E. J., 1991. Conditions for the
equivalence of continuous to pulsed low dose-rate brachytherapy. International Journal of Radiation Oncology, Biology, Physics, 20, 181-190. BROERSE, J. J., MIJNHEER, B. J. & WILLIAM, J. R., 1981.
European protocol for neutron dosimetry for external beam therapy. British Journal of Radiology, 54, 882-898. CATTERALL, M., 1974. A report on three years' fast neutron therapy from the Medical Research Council's cyclotron at Hammersmith Hospital, London. Cancer, 34, 91-95. CATTERALL, M., 1977. The results of randomized and other clinical trials of fast neutrons from the Medical Research Council Cyclotron, London. International Journal of Radiation Oncology, Biology, Physics, 3, 247-253. The British Journal of Radiology, January 1992
Neutron inter comparisons CROUT, N. M. J., FLETCHER, J. G., GREEN, S., SCOTT, M. C. &
TAYLOR, G. C , 1991. In situ neutron spectrometry to 60 MeV in a water phantom exposed to a cancer therapy beam. Physics in Medicine and Biology, 36, 507-519.
HALL, E. J., ZAIDER, M., BIRD, R. & ROBERTS, W., 1983.
Radiobiological studies with therapeutic neutron beams generated by p + - > B e or d + - » B e . British Journal of Radiology, 56, 349-350.
FOWLER, J. F., MORGAN, R. L. & WOOD, C. A. P., 1963.
HORNSEY, S., MYERS, R., PARNELL, C. J., BONNETT, D. E.,
Pre-therapeutic experiments with the fast neutron beam from the Medical Research Council cyclotron. I. The biological and physical advantages and problems of neutron therapy. British Journal of Radiology, 36, 77-80.
biological effectiveness with depth of the Clatterbridge neutron beam. British Journal of Radiology, 61, 1058-1062.
GREENWOOD, L. R., HEINRICH, R. R., SALTMARSH, M. J. &
FULMER, C. B., 1979. Integral tests of neutron activation cross sections in a 9Be(d,n) field at Ed = 40 MeV. Nuclear Science and Engineering, 72, 175-190. GRIFFIN, T., BLASKO, J. & LARAMORE, G., 1979. Results of fast
neutron beam radiotherapy pilot studies at the University of Washington. In High LET Radiations in Clinical Radiotherapy (Supplement to European Journal of Cancer), 23-30. HALL, E. J., 1977. Radiobiological intercomparison in vivo and in vitro. International Journal of Radiation Oncology, Biology, Physics 3, 195-201. HALL, E. J. & KELLERER, A., 1979. Review of RBE data for
cells in culture. In High LET Radiations in Clinical Radiotherapy (Supplement to European Journal of Cancer), 171-174. HALL, E. J., ROIZIN-TOWLE, L., THEUS, R. B. & AUGUST, L. S.,
1975. Radiobiological properties of high energy cyclotron produced neutrons used for radiotherapy. Radiology, 117, 173-178. HALL, E. J., WITHERS, H. R., GERACI, J. P., MEYN, R. E., RASEY, J., TODD, P. & SHELINE, G. E., 1979. Radiobiological
intercomparisons of fast neutron beams used for therapy in Japan and the United States. International Journal of Radiation Oncology, Biology, Physics, 5, 227-233. HALL, E. J., ZAIDER, M., BIRD, R. & ASTOR, M., 1982.
Radiobiological studies with therapeutic neutron beams generated by p + -> Be or d + - » B e . British Journal of Radiology, 55, 640-644.
Vol. 65, No. 769
BLAKE, S. W. & BEWLEY, D. K., 1988. Changes in relative
HORTON J. L., OTTE, V. A., SCHULTEISS, T. E., STAFFORD,
P. M., SUN, T. & ZERMENO, A., 1988. Physical characteristics of the M. D. Anderson Hospital clinical neutron beam. Radiotherapy and Oncology, 13, 17-22. MIJNHEER, B. J., ZOETLIEF, J. & BROERSE, J. J., 1978. Build-up
and depth dose characteristics of different fast neutron beams relevant to radiotherapy. British Journal of Radiology, 51, 122-126. SHELINE, G. E., PHILLIPS, T. L., FIELD, S. B., BRENNAN, J. T. &
RAVENTOS, A., 1971. Effects of fast neutrons on human skin. American Journal of Roentgenology, 111, 31-41. SMITH, A. R., ALMOND, P. R., SMATHERS, J. B., OTTE, V. A., ATTIX, F. H., THEUS, R. B., WOOTTON, P., BICHSEL, H., EENMAA, J., WILLIAMS, D., BEWLEY, D. K. & PARNELL, C. J.,
1975. Dosimetry intercomparisons between fast-neutron radiotherapy facilities. Medical Physics, 2, 195-200. STONE, R. S., 1948. Neutron therapy and specific ionizations. American Journal of Roentgenology, 59, 771-779. TODD, P., GERACI, J. P., FURCINITTI, P. S., Rossi, R. M., MIKAGE, F., THEUS, P. & SCHROY, C. B., 1975. Comparison
of the effects of various cyclotron-produced fast neutrons on the reproductive capacity of cultured human kidney (T-l) cells. International Journal of Radiation Oncology, Biology, Physics, 4, 1015-1022. TSUNEMOTO, H., UMEGAKI, Y., KUTSUTANI, Y., ARAI, T., MORITA, S., KURISU, A., KAWASHIMA, K. & MARUYAMA, T.,
1979. Results of clinical applications of fast neutrons in Japan. In High LET Radiations in Clinical Radiotherapy (Supplement to European Journal of Cancer), 75-78.
71
