Characteristics of neutron beam generated by 500 MeV proton beam

A neutron irradiation facility was constructed at PARMS, University of Tsukuba to produce an ultrahigh energy neutron beam with a depth dose distribution superior to an x-ray beam generated by a modern linac. This neutron beam was produced from the reaction on a thick uranium target struck by a 500 MeV proton beam from the booster synchrotron of the High Energy Physics Laboratory. The percentage depth dose of this neutron beam was nearly equivalent to that of x-rays around 20 MV and the dose rate was 15 cGy per minute. The relative biological effectiveness (RBE) of this neutron beam has been estimated using the cell inactivation effect and the HMV-I cell line. The survival curve of cells after neutron irradiation has a shoulder with n and Dq of 8 and 2.3 Gy, respectively. The RBE value at the 10(-2) survival level for the present neutron beam as compared with 137Cs gamma rays was 1.24. The results suggest that the biological effects of ultrahigh energy neutrons are not large enough to be useful, although the depth dose distribution of neutrons can be superior to that of high energy linac x-rays.     

Changes in RBE of 14-MeV (d + T) neutrons for V79 cells irradiated in air and in a phantom: is RBE enhanced near the surface?

PURPOSE: The relative biological effectiveness (RBE) for inactivation of V79 cells was determined as function of dose at the Heidelberg 14-MeV (d + T) neutron therapy facility after irradiation with single doses in air and at different depths in a therapy phantom. Furthermore, to assess the reproducibility of RBE determinations in different experiments we examined the relationship between the interexperimental variation in radiosensitivity towards neutrons with that towards low LET 60Co photons. METHODS: Clonogenic survival of V79 cells was determined using the colony formation assay. The cells were irradiated in suspension in small volumes (1.2 ml) free in air or at defined positions in the perspex phantom. Neutron doses were in the range, Dt = 0.5-4 Gy. 60Co photons were used as reference radiation. RESULTS: The radiosensitivity towards neutrons varied considerably less between individual experiments than that towards photons and also less than RBE. However, the mean sensitivity of different series was relatively constant. RBE increased with decreasing dose per fraction from RBE = 2.3 at 4 Gy to RBE = 3.1 at 0.5 Gy. No significant difference in RBE could be detected between irradiation at 1.6 cm and 9.4 cm depth in the phantom. However, an approximately 20% higher RBE was found for irradiation free in air compared with inside the phantom. Combining the two effects, irradiation with 0.5 Gy free in air yielded an approximately 40% higher RBE than a dose of 2 Gy inside the phantom. CONCLUSION: The measured values of RBE as function of dose per fraction within the phantom is consistent with the energy of the neutron beam. The increased RBE free in air, however, is greater than expected from microdosimetric parameters of the beam and may be due to slow recoil protons produced by interaction of multiply scattered neutrons or to an increased contribution of alpha particles from C(n, alpha) reactions near the surface. An enhanced RBE in subcutaneous layers of skin combined with an increase in RBE at low doses per fraction outside the target volume could potentially have significant consequences for normal tissue reactions in radiotherapy patients treated with fast neutrons.     

Microdosimetric specification of the radiation quality of a d(48.5)+Be fast neutron therapy beam produced by a superconducting cyclotron

The proportional counter microdosimetric technique has been employed to quantify variations in the quality of a d(48.5)+Be fast neutron beam passing through a homogeneous water phantom. Single event spectra have been measured as a function of spatial location in the water phantom and field size. The measured spectra have been separated into component spectra corresponding to the gamma, recoil proton and alpha plus heavy recoil ion contribution to the total absorbed dose. The total absorbed dose normalized to the "monitor units" used in daily clinical use has been calculated from the measured spectra and compared to the data measured with calibrated ion chambers. The present measurements agree with the ion chamber data to within 5%. The RBE of the neutron beam is assumed to be proportional to the microdosimetric parameter y* for the dose ranges pertinent to fractionated neutron therapy. The relative variations in y*, assumed to be representative of variations in the RBE are mapped as a function of field size and spatial location in the phantom. A variation in the RBE of about 4% for points within and 8% for points outside a 10 cm x 10 cm field is observed. The variations in the RBE within the beam are caused by an increase in the gamma component with depth. An increase in the RBE of about 4% is observed with increasing field size which is attributed to a change in the neutron spectrum. Compared to the uncertainties in the prescribed dose, associated with uncertainties in the clinically used RBE, variation in the RBE between various tissues, and other dosimetric uncertainties caused by factors such as patient inhomogeneities, patient setup errors, patient motion, etc., the measured spatial RBE variations are not considered significant enough to be incorporated into the treatment planning scheme.     

Neutron induced recoil protons of restricted energy and range and biological effectiveness

Low energy neutrons (<2 MeV), those of principal concern in radiation protection, principally initiate recoil protons in biological tissues. The recoil protons from monoenergetic neutrons form rectangular distributions with energy. Monoenergetic neutrons of different energies (<2 MeV) will then produce overlapping recoil proton spectra. By overlapping the effects of individual deposition events, determined microdosimetrically for cell nuclear dimensions, from such neutron beams the biological effectiveness of recoil protons within defined energy and range bounds can be determined. Here chromosomal aberrations per cell have been quantified following irradiation of Vicia faba cells with monoenergetic neutrons of 230, 320, 430, and 1,910 keV. Aberration frequencies from cells from part of the cell cycle, thereby limiting nuclear dimensions, were linearly related to dose and to the frequency of proton recoils per nucleus. The 320 keV neutrons were the most biologically effective per unit absorbed dose and 430 keV neutrons most effective per recoil proton, with 21% of recoils inducing aberrations. After extraction of effectiveness per proton recoil within each energy and range bounds (0-230, 230-320, 320-430, and 430-1,910 keV), it was concluded that recoil protons with energies of about 200-300 keV, traveling 2.5-4 microm and depositing energy at about 80 keV micrometer(-1), are more efficient at aberration induction than those recoil protons of lesser range though near equivalent LET and those of greater range through lesser LET. This approach allows for assessment of the biological effectiveness of individual energy deposition events from low energy neutrons, the lowest dose a cell can receive, and provides an alternative to considerations of relative biological effectiveness.     

A calculation of the relative biological effectiveness of 125I and 103Pd brachytherapy sources using the concept of proximity function

The clinical application of encapsulated radioactive sources in brachytherapy plays an important role in the treatment of malignancy. 125I and 103Pd sources have been widely used in the permanent implant of prostate cancer. An important consideration for the choice of brachytherapy sources is their relative biological effectiveness (RBE). Previous calculations of this quantity have used the dose-averaged lineal energy, yD, as a measure of biological effectiveness. In this approach, however, the selection of a relevant site size remains an open question. Here we avoid this problem by using the generalized theory of dual radiation action to calculate the initial slope, alpha, of the dose-effect curves using the proximity function, t(x), and the biological response function, gamma(x). At low doses and/or low dose rates (e.g., prostate implants) the parameter alpha determines the RBE. Proximity function, t(x), is the probability distribution function of distances between pairs of sublesions; and the biological function, gamma(x), is the probability that two sublesions at a distance x apart results in a lesion. Functions t(x) have been calculated for each source using the Monte Carlo transport codes PHOEL and PROTON5. The function gamma(x) has been taken from a published analysis. The RBE values thus obtained are: 1.5 for 125I and 1.6 for 103Pd. The question of whether an "effective" site size exists where yD approximates best the variation of alpha with radiation quality is also addressed.     

Neutron radiobiology revisited

The present paper reviews the experimental results of normal tissue and tumour studies in animals. The dose per fraction dependence of the RBE in normal tissues has been long recognised, together with the steeper increase of RBE at low doses for late responding tissues compared with acute reactions. The dose dependence for tumours is more complex, because of hypoxia and reoxygenation, as well as differences in repair capability after high LET damage. A comparison of tumour and normal tissue RBE values shows that there is little experimental evidence for a therapeutic advantage at clinically relevant doses. In particular, the RBE for slow growing tumours is even lower than that for the faster growing mouse tumours. The reasons for the loss of expected neutron benefits in clinically relevant experiments are discussed. The disappointing prospects for neutrons are contrasted with the current multifactorial approaches to overcoming resistance to more conventional low LET radiations, including acceleration, hyperfractionation and several types of hypoxic cell radiosensitizers.     

The intraoperative administration of ketorolac tromethamine in evaluating length of stay in a same day surgery unit

Relative biological effectiveness (RBE), as a function of linear energy transfer (LET), is evaluated for different types of damage contributing to mammalian cell reproductive death. Survival curves are analysed assuming a linear-quadratic dose dependence of lethal lesions. The linear term represents lethal damage due to single particle tracks, the quadratic term represents lethality due to interaction of lesions from independent tracks. RBE-LET relationships of single-track lethal damage, sublethal damage, potentially lethal damage and DNA double-strand breaks (dsb) are compared. Single-track lethal damage is shown to be composed of two components: damage that remains unrepaired in an interval between irradiation and assay, characterized by a very strong dependence on LET, with RBEs up to 20, and potentially lethal damage, which is weakly dependent on LET with RBEs < 3. Potentially lethal damage and sublethal damage depend similarly on LET as DNA dsb. The identification of these different components of damage leads to an interpretation of differences in radiosensitivity and in RBEs among various types of cells.     

Chromosome aberrations in human fibroblasts induced by monoenergetic neutrons. I. Relative biological effectiveness

The relative biological effectiveness (RBE) of neutrons for many biological end points varies with neutron energy. To test the hypothesis that the RBE of neutrons varies with respect to their energy for chromosome aberrations in a cell system that does not face interphase death, we studied the yield of chromosome aberrations induced by monoenergetic neutrons in normal human fibroblasts at the first mitosis postirradiation. Monoenergetic neutrons at 0.22, 0.34, 0.43, 1, 5.9 and 13.6 MeV were generated at the Accelerator Facility of the Center for Radiological Research, Columbia University, and were used to irradiate plateau-phase fibroblasts at low absorbed doses from 0.3 to 1.2 Gy at a low dose rate. The reference low-LET, low-dose-rate radiation was 137Cs-gamma rays (0.66 MeV). A linear dose response (Y = alphaD) for chromosome aberrations was obtained for all monoenergetic neutrons and for the gamma rays. The yield of chromosome aberrations per unit dose was high at low neutron energies (0.22, 0.34 and 0.43 MeV) with a gradual decline with the increase in neutron energy. Maximum RBE (RBEm) values varied for the different types of chromosome aberrations. The highest RBE (24.3) for 0.22 and 0.43 MeV neutrons was observed for intrachromosomal deletions, a category of chromosomal change common in solid tumors. Even for the 13.6 MeV neutrons the RBEm (11.1) exceeded 10. These results show that the RBE of neutrons varies with neutron energy and that RBEs are dissimilar between different types of asymmetric chromosome aberrations and suggest that the radiation weighting factors applicable to low-energy neutrons need firmer delineation. This latter may best be attained with neutrons of well-defined energies. This would enable integrations of appropriate quality factors with measured radiation fields, such as those in high-altitude Earth atmosphere. The introduction of commercial flights at high altitude could result in many more individuals being exposed to neutrons than occurs in terrestrial workers, emphasizing the necessity for better-defined estimates of risk.     

Comparative evaluation of markers of bone resorption in patients with breast cancer-induced osteolysis before and after bisphosphonate therapy

The induction of the his(-)-->his+ mutants in vegetative and spores of Bacillus subtilis wild-type cells irradiated with gamma rays and helium ions (LET = 20-80 keV/micron) has been investigated. It was shown that the dose dependence of the mutation induction in vegetative cells is described by a linear-quadratic function of dose in case of both gamma-rays and helium ions. RBE (LET) dependencies on the lethal and mutagenic effect of irradiation have a local maximum. The maximum of RBE (LET) dependence on the mutagenic assay is shifted at the low region of LET in comparison with the lethal effect of irradiation.     

Monte Carlo dosimetry study of a 6 MV stereotactic radiosurgery unit

Small-field and stereotactic radiosurgery (SRS) dosimetry with radiation detectors, used for clinical practice, have often been questioned due to the lack of lateral electron equilibrium and uncertainty in beam energy. A dosimetry study was performed for a dedicated 6 MV SRS unit, capable of generating circular radiation fields with diameters of 1.25-5 cm at isocentre using the BEAM/EGS4 Monte Carlo code. With this code the accelerator was modelled for radiation fields with a diameter as small as 0.5 cm. The radiation fields and dosimetric characteristics (photon spectra, depth doses, lateral dose profiles and cone factors) in a water phantom were evaluated. The cone factor (St) for a specific cone c at depth d is defined as St(d, c) = D(d, c)/D(d, c(ref)), where c(ref) is the reference cone. To verify the Monte Carlo calculations, measurements were performed with detectors commonly used in SRS such as small-volume ion chambers, a diamond detector, TLDs and films. Results show that beam energies vary with cone diameter. For a 6 MV beam, the mean energies in water at the point of maximum dose for a 0.5 cm cone and a 5 cm cone are 2.05 MeV and 1.65 MeV respectively. The values of St obtained by the simulations are in good agreement with the results of the measurements for most detectors. When the lateral resolution of the detectors is taken into account, the results agree within a few per cent for most fields and detectors. The calculations showed a variation of St with depth in the water. Based on calculated electron spectra in water, the validity of the assumption that measured dose ratios are equal to measured detector readings was verified.     

Initial recombination in a parallel-plate ionization chamber exposed to heavy ions

For exact determination of absorbed dose in heavy-ion irradiation fields which are used in radiation therapy and biological experiments, ionization chambers have been characterized with defined heavy-ion beams and correction factors. The LET (linear energy transfer) dependence of columnar recombination in a parallel-plate ionization chamber has been examined. Using 135 MeV/u carbon and neon beams, the ion collection efficiency was measured for several gases (air, carbon dioxide, argon and tissue-equivalent gas). 95 MeV/u argon beams and 90 MeV/u iron beams were also used for measurements of columnar recombination in air. As expected by Jaffe theory, the inverse of the ratio of the ionization charge to the saturated ionization charge had a linear relationship with the inverse of the electric field strength in the region below 0.002 V(-1) cm. The gradient of the line increases as the LET of the heavy ions increases. A strong LET dependence of the gradient was observed in air and carbon dioxide. The LET dependence was not observed in tissue-equivalent gas, nitrogen or argon. The exact depth-dose distribution of the heavy-ion beam was obtained by this correction of the initial recombination effect for the collected ionization charge. The columnar recombination in air was analysed using Jaffe theory; the obtained parameter b (a track radius) should be in the range between 0.001 cm and 0.005 cm, whereas the value obtained by Jaffe is 0.00179 cm. The value of the parameter b should increase as the LET of the heavy-ion beam increases in order to reproduce the experimental values of the initial recombination.     

Substructure in the cell survival response at low radiation dose: effect of different subpopulations

In order to obtain more accurate measurements of cell survival after low doses of radiation, we have used the cell sorter assay, in which a cell sorter is used to accurately count out the number of cells plated for colony formation. This method, combined with data averaging, permits measurements of survival with superior precision, which have revealed that there is substructure in the radiation response of asynchronously dividing Chinese hamster cells. The substructure, observed at doses of a few Gy, has features of a 2-component response, consistent with the presence of subpopulations of cells of different cell-cycle-related radiosensitivity. The absence of any substructure in the radiation response of homogeneous (tightly synchronized) cell populations lends strong support to this subpopulation explanation of the substructure. This assay has also been used on a variety of human tumour cell lines, most of which exhibited substructure similar to that of Chinese hamster cells. This paper outlines the application of the cell sorter assay to three different problems: (i) radiosensitizer mechanisms-etanidazole and RB 6145 are shown to enhance primarily the beta term and alpha term, respectively, of tumour cell kill, indicating that sensitizer efficacy may be tumour-specific and predictable from tumour response parameters; (ii) accurate measurement of Relative Biological Effectiveness (RBE) in a modulated clinical proton beam shows that the RBE is both dose- and depth-dependent; and (iii) measurements at lower doses clearly demonstrate a second order of substructure, termed the hypersensitive response, at doses < 1 Gy.     

Dose and LET distributions due to neutrons and photons emitted from stopped negative pions

Computer calculations are made of the dose and LET distributions due to neutrons and photons produced when negative pions are stopped in a phantom. When negative pions are stopped in a material they undergo nuclear capture, resulting in the disintegration of the nucleus and the emission of short range charged particles and longer range neutrons and photons. The uncharged radiation constitutes a potentially large source of dose outside the treatment volume. A simple phantom consisting of a 0-25 m cube of either tissue or bone-equivalent material is set up with a 0-05 m cube in the centre to represent the treatment volume. Neutrons and photons are started in this central volume and transported across the phantom using Monte Carlo transport codes. Several different initial energy spectra for the neutrons are used, taken from experimental and theoretical data. These different spectra are found to give significant differences in dose, though the distance to the 80% dose level is always about 0-015 m. Order of magnitude differences in some LET regions are also found. The dose deposited by neutrons in bone is about 24% less than in soft tissue, the photon dose being small compared with the neutron dose.     

On compensator design for photon beam intensity-modulated conformal therapy

Recently the compensator has been shown to be an in expensive and reliable dose delivery device for photon beam intensity-modulated radiation therapy (IMRT). The goal of IMRT compensator design is to produce an optimized primary fluence profile at the patient's surface obtained from the optimization procedure. In this paper some of the problems associated with IMRT compensator design, specifically the beam perturbations caused by the compensator, are discussed. A simple formula is derived to calculate the optimal compensator thickness profile from an optimized primary fluence profile. The change of characteristics of a 6 MV beam caused by the introduction of cerrobend compensators in the beam is investigated using OMEGA Monte Carlo codes. It is found that the compensator significantly changes the energy spectrum and the mean energy of the primary photons at the patient's surface. However, beam hardening does not have as significant an effect on the percent depth dose as it does on the energy spectrum. We conclude that in most situations the beam hardening effect can be ignored during compensator design and dose calculation. The influence of the compensator on the contaminant electron buildup dose is found to be small and independent of the compensator thickness of interest. Therefore, it can be ignored in the compensator design and included as a correction into the final dose distribution. The scattered photons from the compensator are found to have no effect on the surface dose. These photons produce a uniform low fluence distribution at the patient's surface, which is independent of compensator shape. This is also true for very large fields and extremely asymmetric and nonuniform compensator thickness profiles. Compared to the primary photons, the scattered photons have much larger angular spread and similar energy spectrum at the patient's surface. These characteristics allow the compensator thickness profile and the dose distribution to be calculated from the optimized fluence profile of primary photons, without considering the scattered photons.     

Experimental tests of proton beam localization

The depth of penetration of heavy charged-particle therapy beams is sensitive to the density of tissues traversed. Maximum depth of dose contours will vary appreciably as the beam passes through bone, muscle, lung, and air or gas. Calculations suggest that beam activation of the short-lived positron-emitting isotope 15O in vivo will permit localization of proton therapy beams with resonable detected-event density and dose. Preliminary tests of this method indicate that the beam can be located at depth with a typical dose of 15 rad, using a large field-of-view positron camera on-ling. This technique is also applicable to other heavy charged-particle beams, negative pions, and heavy ions.     

Radiobiology of alpha particles. IV. Cell inactivation by alpha particles of energies 0.4-3.5 MeV

In this report the effectiveness of low-energy alpha particles in the range 0.4 to 3.5 MeV for cell killing is investigated. Four cell lines of different nuclear dimensions (AG1522, C3H 10T1/2, CHO-10B, and HS-23) are studied. Monte Carlo simulations are carried out to interpret the experimental results. They are presented as a function of dose to the nucleus, the total track length of alpha particles in the nucleus, and other parameters. It is found that the effectiveness of alpha particles for cell killing decreases with decreasing alpha-particle energy. The maximum RBE value is found to extend to LET values as high as 180 keV/microns. Although the LET might be the same, the effectiveness of alpha particles for cell killing is higher in the ascending part of the Bragg curve compared to descending part of the Bragg curve. The terminal tracks of alpha particles are observed to be less effective for cell killing.     

Response analysis of TLD-300 dosimeters in heavy-particle beams

In vivo dosimetry is recommended as part of the quality control procedure for treatment verification in radiation therapy. Using thermoluminescence, such controls are planned in the p(65) + Be neutron and 85 MeV proton beams produced at the cyclotron at Louvain-La-Neuve and dedicated to therapy applications. A preliminary study of the peak 3 (150 degrees C) and peak 5 (250 degrees C) response of CaF2:Tm (TLD-300) to neutron and proton beams aimed to analyse the effect of different radiation qualities on the dosimetric behaviour of the detector irradiated in phantom. To broaden the range of investigation, the study was extended to an experimental 12C heavy ion beam (95 MeV/nucleon). The peak 3 and 5 sensitivities in the neutron beam, compared to 60Co, varied little with depth. A major change of peak 5 sensitivity was observed for samples positioned under five leaves of the multi-leaf collimator. While peak 3 sensitivity was constant with depth in the unmodulated proton beam, peak 5 sensitivity increased by 15%. Near the Bragg peak, peak 3 showed the highest decrease of sensitivity. In the modulated proton beam, the sensitivity values were not significantly smaller than those measured in the unmodulated beam far from the Bragg peak region. The ratio of the heights of peak 3 and peak 5 decreased by 70% from the 60Co reference radiation to the 12C heavy-ion beam. This parameter was strongly correlated with the change of radiation quality.     

Dose perturbation caused by high-density inhomogeneities in small beams in stereotactic radiosurgery

The influence of high-density tissue heterogeneities in small-diameter beams used in stereotactic radiosurgery has been investigated. Dose perturbation immediately behind aluminium sheets, used to simulate a high-density tissue inhomogeneity such as bone, was studied in a solid water phantom. Dose reduction factors (DRFs), which are the ratios of the dose in the presence of the inhomogeneity to dose in a uniform density solid water phantom, were measured with a diamond detector for three thicknesses of aluminium. DRFs exhibit dependence on both the inhomogeneity thickness and the beam diameter. The DRF decreases with inhomogeneity thickness. The DRF initially decreases with increase in the beam diameter from 12.5 to 25 mm. For fields greater than 25 mm, the DRFs are nearly constant. The commonly used algorithms such as the TAR ratio method underestimate the magnitude of the measured effect. A good agreement between these measurements and Monte Carlo calculations is obtained. The influence of the high-density inhomogeneity on the tissue maximum ratio (TMR) was also measured with the inhomogeneity at a fixed depth dmax from the entrance surface. The TMR is reduced for all detector-inhomogeneity distances investigated. The dose build-up phenomenon observed in the presence of low-density air inhomogeneity is absent in the presence of a high-density inhomogeneity. The beam width (defined by 50% dose points) immediately beyond the inhomogeneity is unaffected by the high-density inhomogeneity. However, the 90%-10% and 80%-20% dose penumbra widths and the dose outside the beam edge (beyond the 50% dose point) are reduced. This reduction in dose outside the beam edge is caused by the reduced range of the secondary radiation (photons and electrons) in the high-density medium.     

A proton dose calculation algorithm for conformal therapy simulations based on Molière's theory of lateral deflections

An algorithm is developed for computing proton dose distributions in the therapeutic energy range (100-250 MeV). The goal is to provide accurate pencil beam dose distributions for two-dimensional or three-dimensional simulations of possible intensity-modulated proton therapy delivery schemes. The algorithm is based on Molière's theory of lateral deflections, which accurately describes the distribution of lateral deflections suffered by incident charged particles. The theory is applied to nonuniform targets through the usual pencil beam approximation which assumes that all protons from a given pencil beam pass through the same material at each depth. Fluence-to-dose conversion is made via Monte Carlo calculated broad-field central-axis depth-dose curves, which accounts for attenuation due to nuclear collisions and range straggling. Calculation speed is enhanced by using a best-fit Gaussian approximation of the radial distribution function at depth. Representative pencil beam and spread-out Bragg-peak computations are presented at 250 MeV and 160 MeV in water. Computed lateral full-widths-at-half-maximum's in water, at the Bragg peak, agree with the expected theoretical lateral values to within 1% at 160 MeV and to within 3% at 250 MeV. This algorithm differs from convolution methods in that the effect of the depth of any inhomogeneities in density or atomic composition are accounted for in a rigorous fashion. The algorithm differs from Fermi-Eyges based methods by accounting in a rigorous way for the effect of nonsmall-angle scattering and screening due to atomic electrons. The computational burden is only slightly greater than that expected using the less-rigorous Fermi-Eyges theory.     

In-phantom dosimetry and spectrometry of photoneutrons from an 18 MV linear accelerator

A combination of three superheated drop detectors with different neutron energy responses was developed to evaluate dose-equivalent and energy distributions of photoneutrons in a phantom irradiated by radiotherapy high-energy x-ray beams. One of the three detectors measures the total neutron dose equivalent and the other two measure the contributions from fast neutrons above 1 and 5.5 MeV, respectively. In order to test the new method, the neutron field produced by the 10 cm X 10 cm x-ray beam of an 18 MV radiotherapy accelerator was studied. Measurements were performed inside a tissue-equivalent liquid phantom, at depths of 1, 5, 10 and 15 cm and at lateral distances of 0, 10, and 20 cm from the central axis. These data were used to calculate the average integral dose to the radiotherapy patient from direct neutrons as well as from neutrons transmitted through the accelerator head. The characteristics of the dosimeters were confirmed by results in excellent agreement with those of prior studies. Track etch detectors were also used and provided an independent verification of the validity of this new technique. Within the primary beam, we measured a neutron entrance dose equivalent of 4.5 mSv per Gy of photons. It was observed that fast neutrons above 1 MeV deliver most of the total neutron dose along the beam axis. Their relative contribution increases with depth, from about 60% at the entrance to over 90% at a depth of 10 cm. Thus, the average energy increases with depth in the phantom as neutron spectra harden.