In vivo assessment of magnesium status in human body using accelerator-based neutron activation measurement of hands: a pilot study.

Magnesium (Mg) is an element essential for many enzymatic reactions in the human body. Various human and animal studies suggest that changes in Mg status are linked to diseases such as cardiac arrhythmia, coronary heart disease, hypertension, premenstrual syndrome, and diabetes mellitus. Thus, knowledge of Mg levels in the human body is needed. A direct measurement of human blood serum, which contains only 0.3% of the total body Mg, is generally used to infer information about the status of Mg in the body. However, in many clinical situations, Mg stored in large levels, for example in bones, muscles, and soft tissues, needs to be monitored either to evaluate the efficacy of a treatment or to study the progression of diseases associated with the deficiency of total body Mg. This work presents a feasibility study of a noninvasive, in vivo neutron activation analysis (IVNAA) technique using the 26Mg (n, gamma) 27Mg reaction to measure Mg levels in human hands. The technique employs the McMaster University high beam current Tandetron accelerator hand irradiation facility and an array of eight NaI (T1) detectors arranged in a 4 pi geometry for delayed counting of the 0.844 and 1.014 MeV gamma rays emitted when 27Mg decays in the irradiated hand. Mg determination in humans using IVNAA of hands has been demonstrated to be feasible, with effective doses as low as one-quarter of those delivered in chest x rays. The overall experimental uncertainty in the measurements is estimated to be approximately 5% (1 sigma). The results are found to be in the range of the in vitro measurements reported for other cortical bones collected from different sites of the human skeleton, which confirms that this technique mainly provides a measure of the amount of Mg in hand bones. The average concentration of Mg determined in human hands is 10.96 +/- 1.25 (+/- 1 SD) mg Mg/g Ca. The coefficient of variation (11%) observed in this study is comparable with or lower than several studies using in vitro measurements reported in the literature and therefore allows for a quantitative intersubject comparison, even if to a limited extent. The features of the developed technique such as its simplicity, rapidity, accuracy, robustness, noninvasive nature, and very effective use of radiation doses, present the technique as a viable diagnostic tool available for trial in a clinical environment.

Dosimetric characterization of the irradiation cavity for accelerator-based in vivo neutron activation analysis.

A neutron irradiation cavity for in vivo activation analysis has been characterized to estimate its dosimetric specifications. The cavity is defined to confine irradiation to the hand and modifies the neutron spectrum produced by a low energy accelerator neutron source to optimize activation per dose. Neutron and gamma-ray dose rates were measured with the microdosimetric technique using a tissue-equivalent proportional counter at the hand irradiation site and inside the hand access hole. For the outside of the cavity, a spherical neutron dose equivalent meter and a Farmer dosemeter were employed instead due to the low intensity of the radiation field. The maximum dose equivalent rate at the outside of the cavity was 2.94 microSv/100 microA min, which is lower by a factor of 1/2260 than the dose rate at the hand irradiation position. The local dose contributions from a hand, an arm and the rest of a body to the effective dose rate were estimated to be 1.73, 0.782 and 2.94 microSv/100 microA min, respectively. For the standard irradiation protocol of the in vivo hand activation, 300 microA min, an effective dose of 16.3 microSv would be delivered.

Monte Carlo simulation of neutron irradiation facility developed for accelerator based in vivo neutron activation measurements in human hand bones.

The neutron irradiation facility developed at the McMaster University 3 MV Van de Graaff accelerator was employed to assess in vivo elemental content of aluminum and manganese in human hands. These measurements were carried out to monitor the long-term exposure of these potentially toxic trace elements through hand bone levels. The dose equivalent delivered to a patient during irradiation procedure is the limiting factor for IVNAA measurements. This article describes a method to estimate the average radiation dose equivalent delivered to the patient's hand during irradiation. The computational method described in this work augments the dose measurements carried out earlier [Arnold et al., 2002. Med. Phys. 29(11), 2718-2724]. This method employs the Monte Carlo simulation of hand irradiation facility using MCNP4B. Based on the estimated dose equivalents received by the patient hand, the proposed irradiation procedure for the IVNAA measurement of manganese in human hands [Arnold et al., 2002. Med. Phys. 29(11), 2718-2724] with normal (1 ppm) and elevated manganese content can be carried out with a reasonably low dose of 31 mSv to the hand. Sixty-three percent of the total dose equivalent is delivered by non-useful fast group (> 10 keV); the filtration of this neutron group from the beam will further decrease the dose equivalent to the patient's hand.

COMPREHENSIVE RADIATION DOSE MEASUREMENTS AND MONTE CARLO SIMULATION FOR THE 7Li(p,n) ACCELERATOR NEUTRON FIELD.

In order to investigate the radiation dose dependence on the incident proton energy, neutron and gamma-ray doses were measured using a tissue-equivalent proportional counter in the proton energy range of 1.95-2.50 MeV for the McMaster 7Li(p,n) neutron facility. Microdosimetric spectra were collected, and absorbed doses were determined at various positions inside the irradiation cavity, along the lateral axis and outside the shield to find out the spatial distributions of neutron and gamma-ray doses for each proton energy. In parallel with the absorbed dose measurements, MCNP Monte Carlo simulations were carried out and neutron fluence spectra were computed at various positions, which enabled determination of the neutron weighting factors. It was found that neutrons make a substantially dominant contribution to the total equivalent dose for most proton energies and positions. The effective dose for a human subject increased from 0.058 to 1.306 µSv µA-1 min-1 with the increase of proton energy from 1.95 to 2.5 MeV. It is expected that the reported data will be useful for 7Li(p,n) accelerator neutron users.

In vivo neutron activation study of the short-term kinetic behaviour of sodium and chlorine in the human hand.

The time-dependent behaviour of sodium and chlorine was studied as a spinoff from a study of aluminum in the hand of subjects suffering from Alzheimer's disease and a control group, involving 15 Alzheimer's and 16 control subjects with an age range of 63-89 years. This was achieved using the in vivo neutron activation analysis system developed at McMaster University for the non-invasive measurement of aluminum, where a subject's hand is placed in a beam of accelerator-based thermalized neutrons, which activates elements by neutron capture. Following irradiation, the subject's hand is placed in a detection system comprising 9 NaI(Tl) detectors arranged in a 4π geometry to measure activated elements. The redistribution half-lives of the activation products 24Na and 38Cl from the hand were determined after correction for the physical half-life, by means of sequential analysis of the residual activity in the hand. The kinetic behaviours of sodium and chlorine were best characterized by an exponential function corresponding to the rapidly exchangeable pool. The mean redistribution half-lives from the hand for sodium and chlorine in the control subjects were 40.5 ± 17.4 min and 24.2 ± 8.5 min, respectively. For Alzheimer's disease subjects the mean redistribution half-lives were 58.2 ± 36.1 min for sodium and 33.6 ± 16.7 min for chlorine. There was no significant difference in chlorine and sodium redistribution half-lives between the Alzheimer's disease and control group subjects. These results are promising, given that the irradiation and counting protocol were optimized for the aluminum study, rendering them suboptimal for analyzing other elements and their rate of change with time. Further improvements include optimizing the irradiation protocol, longer counting times, and measuring the activity in the un-irradiated hand in various time intervals following irradiation.

Optimization of data analysis for the in vivo neutron activation analysis of aluminum in bone.

An existing system at McMaster University has been used for the in vivo measurement of aluminum in human bone. Precise and detailed analysis approaches are necessary to determine the aluminum concentration because of the low levels of aluminum found in the bone and the challenges associated with its detection. Phantoms resembling the composition of the human hand with varying concentrations of aluminum were made for testing the system prior to the application to human studies. A spectral decomposition model and a photopeak fitting model involving the inverse-variance weighted mean and a time-dependent analysis were explored to analyze the results and determine the model with the best performance and lowest minimum detection limit. The results showed that the spectral decomposition and the photopeak fitting model with the inverse-variance weighted mean both provided better results compared to the other methods tested. The spectral decomposition method resulted in a marginally lower detection limit (5µg Al/g Ca) compared to the inverse-variance weighted mean (5.2µg Al/g Ca), rendering both equally applicable to human measurements.

In vivo measurement of bone aluminium: recent developments.

A biomarker of aluminium accumulation in the human body can play a valuable role in determining health effects of chronic aluminium exposure, complementing other human and environmental monitoring data. In vivo neutron activation provides such a non-invasive biomarker. To date, the best in vivo neutron activation system used thermalised neutrons from a nuclear reactor at Brookhaven National Laboratory, which suffered only slightly from interference from other elements, primarily phosphorus, and from the disadvantage of restricted accessibility. At McMaster, we use a nuclear reaction on an accelerator to select neutron energy, which eliminates the interferences. Spectral decomposition analysis improved sensitivity. A new 4pi detection system also enhanced sensitivity. Together these improvements yield a minimum detection limit of 0.24 mgAl in a hand, slightly better than at Brookhaven and equivalent to "normal" levels. Further improvements should result from a new irradiation cavity and from using a higher proton current on the accelerator to shorten irradiation times. The system is now ready for pilot human studies.

Feasibility of measuring selenium in humans using in vivo neutron activation analysis.

Selenium (Se) is an element that, in trace quantities, plays an important role in the normal function of a number of biological processes in humans. Many studies have demonstrated that selenium deficiency in the body may contribute to an increased risk for certain neoplastic, cardiovascular, osseous, and nervous system diseases including retardation of bone formation. However, at higher concentrations Se is cytotoxic. For these reasons it is desirable to have a means of monitoring selenium concentration in humans.This paper presents the outcome of a feasibility study carried out for measuring selenium in humans using in vivo neutron activation analysis (IVNAA). In this technique a small dose of neutrons is delivered to the organ of interest, the neutrons are readily captured by the target nuclei, and the γ-rays given off are detected outside of the body. For the present study, human hand (bone) tissue equivalent phantoms were prepared with varying amounts of Se. These were irradiated by a low energy fast neutron beam produced by the (7)Li(p,n)(7)Be reaction employing the high beam current Tandetron accelerator. The counting data saved using a 4π NaI(TI) detection system were analyzed. The selenium was detected via the neutron capture reaction, (76)Se(n,γ)(77 m)Se, whereas calcium was detected through the (48)Ca(n,γ)(49)Ca reaction for the purpose of normalization of the Se signals to the calcium signals. From the calibration lines drawn between Se/Ca concentrations and Se/Ca counts ratio, the minimum detection limits (MDLs) were computed for two sets of phantoms irradiated under different irradiation parameters.In this study the optimized MDL value was determined to be 81 ng g(-1) (Se/phantom mass) for an equivalent dose of 188 mSv to the phantom. This MDL was found at least 10 times lower than the reported data on Se concentration measured in bone tissues. It was concluded that the NAA technique would be a feasible means of performing in vivo measurements of selenium in humans. Currently the data on in vivo measurement of selenium in humans are limited; the results of the present study would greatly contribute to the present data.

A neutron activation technique for manganese measurements in humans.

Manganese (Mn) is an essential element for humans, animals, and plants and is required for growth, development, and maintenance of health. Studies show that Mn metabolism is similar to that of iron, therefore, increased Mn levels in humans could interfere with the absorption of dietary iron leading to anemia. Also, excess exposure to Mn dust, leads to nervous system disorders similar to Parkinson's disease. Higher exposure to Mn is essentially related to industrial pollution. Thus, there is a benefit in developing a clean non-invasive technique for monitoring such increased levels of Mn in order to understand the risk of disease and development of appropriate treatments. To this end, the feasibility of Mn measurements with their minimum detection limits (MDL) has been reported earlier from the McMaster group. This work presents improvement to Mn assessment using an upgraded system and optimized times of irradiation and counting for induced gamma activity of Mn. The technique utilizes the high proton current Tandetron accelerator producing neutrons via the (7)Li(p,n)(7)Be reaction at McMaster University and an array of nine NaI (Tl) detectors in a 4 π geometry for delayed counting of gamma rays. The neutron irradiation of a set of phantoms was performed with protocols having different proton energy, current and time of irradiation. The improved MDLs estimated using the upgraded set up and constrained timings are reported as 0.67 µgMn/gCa for 2.3 MeV protons and 0.71 µgMn/gCa for 2.0 MeV protons. These are a factor of about 2.3 times better than previous measurements done at McMaster University using the in vivo set-up. Also, because of lower dose-equivalent and a relatively close MDL, the combination of: 2.0 MeV; 300 µA; 3 min protocol is recommended as compared to 2.3 MeV; 400 µA; 45 s protocol for further measurements of Mn in vivo.

Measurements of fluorine in contemporary urban Canadians: a comparison of the levels found in human bone using in vivo and ex vivo neutron activation analysis.

Non-invasive in vivo neutron activation analysis (NAA) was used to measure the fluorine concentration in 35 people in Hamilton, Ontario, Canada. Measurement and precision data of this second generation NAA system were determined in 2013, and the results were compared with the performance of a first generation system used in a pilot study of 33 participants from the Hamilton area in 2008. Improvements in precision in line with those predicted by phantom studies were observed, but the use of fewer technicians during measurement seemed adversely to affect performance. We compared the levels of fluorine observed in people between the two studies and found them to be comparable. The average fluorine concentration in bone was found to be 3 ± 0.3 mg and 3.5 ± 0.4 mg F/g Ca for 2013 and 2008 measurements respectively. Ten people were measured in both studies; the observed average change in bone fluorine in this subgroup was consistent with that predicted by the observation of the relationship between bone fluorine and age in the wider group. In addition, we observed differences in the relationship between bone fluorine level and age between men and women, which may be attributable either to sex or gender differences. The rate of increase in fluorine content for men was found to be 0.096 ± 0.022 mg F/g Ca per year while the rate of increase for women was found to be slightly less than half that of men, 0.041 ± 0.017 mg F/g Ca per year. A discontinuity in the rate of increase in fluorine content with age was observed in women at around age 50. Bone fluorine content was significantly lower ([Formula: see text]) in women age 50 to 59 than in women age 40 to 49, which we suggest may be attributable to bone metabolism changes associated with menopause. We also observed increased fluorine levels in tea drinkers as compared to non-tea drinkers, suggesting tea may be a significant source of exposure in Canada. The rate of increase in fluorine content of the tea drinkers and the non-tea drinkers were found to be 0.127 (± 0.029) and 0.050 (± 0.009) mg F/g Ca per year respectively. Finally, we also obtained twelve bone samples from cadavers' skulls. Neutron activation analysis was used to determine the fluorine levels in these ex vivo samples. The rate of increase of fluorine content versus age for in vivo and ex vivo measurements were found to be 0.078 ± 0.014 and 0.078 ± 0.050 mg F/g Ca per year respectively. Excellent agreement was found between the fluorine levels determined in vivo and ex vivo using the two separate systems, providing confidence in the fluorine concentration data being measured in vivo.

Beta dose point kernels for radionuclides of potential use in radioimmunotherapy.

Beta dose point kernels for 32P, 67Cu, 90Y, 131I, 186Re, and 188Re nuclides appropriate for radioimmunotherapy are calculated based upon Monte Carlo results. The calculations are shown to differ significantly from values based upon solutions to the electron transport equation. Agreement with experiment for 32P is found to be improved for the former as compared with the latter. Values of the scaled dose point kernels are tabulated at 4% intervals of the continuous slowing down approximation range for each of the six radionuclides. Beta dose distributions are also tabulated at corresponding distances from the source. This data may be used to calculate the spatial dose distribution expected following administration of radiolabeled monoclonal antibodies, aiding in optimum selection of the appropriate radionuclide. Parameters for functions providing analytic representation of the calculated scaled dose point kernels of individual beta groups are presented.

The influence of track structure on the understanding of relative biological effectiveness for induction of chromosomal exchanges in human lymphocytes.

A biophysical model has been applied to describe the production of exchange chromosomal aberrations (dicentrics) in human lymphocytes by radiations of different qualities. The model includes a detailed description of the energy deposition pattern in the form of computer-generated tracks. Energy deposition events are further converted to DNA double-strand breaks (DSBs). Formation of chromosomal exchanges is modeled in competition with repair in a distance-dependent manner with breaks in proximity being most likely to interact. We demonstrate that an assumption of an RBE > 1 for production of DSBs at higher LET leads to a significant increase with LET of both the linear and the quadratic coefficients of the dose response for exchange formation. The latter is not supported experimentally and argues against high RBE values for production of DSBs, at least for those breaks involved in chromosomal exchanges. Assuming that the RBE for production of DSBs is unity, the calculated dose-response curves conformed to experimental data for 60Co gamma rays, 250 kVp X rays and 8.7 MeV protons. The linear coefficient for 23.5 MeV 3He ions is underpredicted. The model predicts that a quadratic term in the dose response for exchange aberrations should be observed at LET values of 20-30 keV/microm. The curvature is not observed experimentally, and the contradiction is discussed.

Gamma ray nuclear resonance absorption: an alternative method for in vivo body composition studies.

We have evaluated gamma ray nuclear resonance absorption (gamma-NRA) on nitrogen, a mature technology proposed and developed by Soreq NRC for detecting explosives, as an alternative to neutron activation for in vivo assaying of body nitrogen. The principles of the gamma-NRA method are outlined, and a test facility constructed at McMaster University's Accelerator Laboratory is described. The results of a feasibility study recently performed there on phantoms and animal tissue are presented and discussed. gamma-NRA is a full imaging technique that essentially constitutes element-specific absorptiometry--i.e., it can generate projections of the mass distribution for a specific element, along with a conventional radiograph of the patient. From the transmission profile of an individual scanned by 9.17 MeV gamma rays, local or whole body nitrogen content can be determined via the resonant attenuation undergone when the beam encounters regions of nitrogen concentration. The advantages of gamma-NRA over neutron activation are (a) radiation doses delivered to the body are at least one order of magnitude lower, thus allowing repeated measurements on individual patients and also rendering the method ethically acceptable for application to children; (b) gamma-NRA is inherently free from uncertainties related to nonuniform distributions of the element in question within the body; (c) it is applicable to patients of varying size and shape; and (d) it yields both nitrogen images and conventional radiographic images of the body.

Analytic representation of the dose from a 32P-coated stent.

The dose along the radial direction located at the midplane of a radioactive stent, simulated by a uniform cylinder of 32P, is represented by an analytical function consisting of the sum of two modified exponentials. This procedure reproduces values obtained from numerical integration, for which no closed form exists, to within 5% for distances up to 6 mm from the wall and for stent diameters from 2-6 mm.

A statistical investigation of the scaling factor method of beta-ray dose distribution derivation: the scaling factor for water to bone.

Reliable methods of estimating doses are essential for the use of beta emitting radionuclides for radiotherapy. The passage of electrons through matter is a very complex phenomenon due to the large number of elastic and inelastic interactions resulting in scattering and energy losses. The analytical solution for the electron transport being intractable, the problem has been addressed by the Monte Carlo technique. Empirical or semiempirical less time consuming methods, such as the scaling factor method, may appear more preferable in practice when dealing with complicated source distributions. The method, proposed by Cross and co-workers [AECL Report Nos. AECL-1617 (1982), AECL 10521 (1992)] consists in the derivation of beta-ray dose distribution in other media from those in water by using a "scaling factor" or "relative attenuation factor" on distance and a closely related renormalization factor imposed by the energy conservation. This work investigates the accuracy of the scaling factor method using a statistical approach, a generalized chi 2 test, focusing on the particular case of potential interest, the scaling factor for water to bone. The direct comparison of the shapes of the depth dose deposition curves in the two media indicates discrepancies of less than 5% up to at least 60% of the range in bone, a depth within which 95% of the initial energy is deposited. The scaling factor derived by this method, 0.9720 +/- 0.0012, confirms the existing experimentally determined value of 0.973 +/- 1% [AECL Report No. AECL-10521 (1992)]. The accuracy of the determination is increased by almost a factor of 10. A way of improving the scaling method, especially for depth over the 60% continuous slowing down approximation range, by using a modulation function is also proposed.

Application of the scaling factor method to estimation of beta dose distributions for dissimilar media separated by a planar interface.

The most accurate method of calculating beta dose distribution currently relies on the Monte Carlo technique. The major drawback of the method is the long computing time required to follow a large number of "electron histories" in order to achieve good statistics, which makes the method unattractive for practical radiation therapy. A way to avoid the Monte Carlo calculations for homogeneous media was suggested by Cross and co-workers (AECL Report Nos. 7617, 1982; 10521, 1992), and is known as the "scaling factor" method. It consists of the determination of the depth dose distribution in a medium based on known data about the dose distribution in an arbitrary reference medium (e.g., air, water) by the use of a scaling factor on distance and a closely related renormalization factor imposed by energy conservation. This work is an attempt to extend the applicability of the scaling factor method to dissimilar media to a planar interface. The investigation was done for an isotropic source of the radioisotope 32P and an interface between water and medium "i," where medium "i" could be any medium with atomic number in the range 8 < Z < 50. The method was checked using three randomly chosen elements 40Zr, 32Ge, and 26Fe, each forming planar interfaces with water at either 100 or 350 mg/cm2. Discrepancies of less than 5% were detected (acceptable for practical radiotherapy) for the depth within which at least 95% of the initial energy is deposited.

Experimental determination of 32P dose backscatter factors at and near soft-tissue boundaries.

Beta-ray dose backscatter factors with respect to soft tissue were measured using an extrapolation chamber. The beta-ray dose backscatter factor is a measure of the change effected in absorbed dose to a soft-tissue medium when part of the medium is replaced by a material other than soft tissue (i.e., a scatterer); the source is located at the boundary between the two media. The dependencies of backscatter factor on scatterer atomic number and on source geometry are investigated, and the variation of backscatter factor with distance from the boundary is determined. For a 32P point source, backscatter factors with respect to Mylar, a soft-tissue substitute, at 0.55 mg/cm2 from the boundary, are, 29.65[0.12]%, 31.07[0.24]%, 19.30[0.48]%, 16.27[0.35]%, 5.46[0.11]% and -26.44[0.02]% for bismuth, tungsten, cadmium, copper, aluminum, and air scatterers, respectively. Backscatter factors measured for a 32P planar source are generally smaller than those for a point source. The variation of backscatter factor with distance from the boundary is well represented analytically by sums of exponentials. Therefore, the rate of decrease of backscatter factor with distance can be specified by a relaxation length, defined as the depth through which the backscatter factor is reduced by 1/e, where e is the base of the natural logarithm. For example, with a 32P planar source, relaxation lengths in Mylar are 588[7] mg/cm2 and 238[2]mg/cm2 for bismuth and aluminum scatterers, respectively. Qualitative interpretation of backscatter factor depth profiles is presented. In addition, the variations of backscatter factor with scatterer atomic number and with source geometry are discussed with reference to existing experimental findings on beta particle reflection.

An evaluation of the EGS4 and CYLTRAN Monte Carlo codes with regard to boundary beta-ray dosimetry by comparison with experimental beta-ray dose backscatter factors.

Beta-ray dose backscatter factors or dose ratios at planar soft-tissue boundaries were calculated using the EGS4/RESTA and CYLTRAN (version 2.1) Monte Carlo codes and these data were compared with experimental results. Since the beta-ray source was 32P, this work addressed the transport of, and energy deposition by, electrons less energetic than 2 MeV. In particular, the simulations targeted the codes' performances with regard to the transport of low energy electrons across material boundaries and the backscattering of low energy electrons. In general, backscatter factors calculated at 7.25 mg/cm2 from several soft-tissue interfaces agreed with experimental values to within about five percent. CYLTRAN was also used to calculate the variation of backscatter factor with distance from aluminum/soft-tissue and air/soft-tissue interfaces and was found to reproduce the shapes of experimental backscatter factor depth profiles.

Effect of tissue inhomogeneity on beta dose distribution of 32P.

In a homogeneous medium of soft tissue the radiation dose distribution due to a nonuniformly distributed beta source can be calculated by convolution of the beta dose point kernel of the nuclide with the source distribution. A possible extension of the technique to the calculation of the dose distribution in heterogeneous media involving relatively simple geometric interfaces requires the knowledge of the resulting perturbation to the beta point kernels in individual media. We simulated a soft-tissue-bone planar interface by a polystyrene (PST)-aluminum junction and measured the change in beta dose from the dose value in homogeneous PST due to a point source of 32P using 7LiF thermoluminescent dosimeters. With the point source at the interface, the dose rates at 0-31, 125-156, and 283-314 mg/cm2 separations from the interface were increased by (12 +/- 3)%, (8 +/- 2)%, and (3 +/- 2)%, respectively, compared with homogeneous PST. With the point source at a PST-air planar interface to simulate a soft-tissue-air junction, the dose rates at 0-31, 139-170, and 283-314 mg/cm2 from the interface were decreased by (25 +/- 4)%, (11 +/- 7)%, and (5 +/- 2)%, respectively. The changes in dose rates for these two interfaces have also been measured with degraded spectra of 32P. Comparison of the experimental data with Monte Carlo calculation for a point source and the two-group method of calculation for a plane source is also presented.

In vivo activation analysis of spinal calcium using 25 keV neutrons.

The applicability of 25 keV neutrons to the in vivo activation analysis of spinal calcium is examined both theoretically and experimentally. It is shown that the use of this energy results in an increase in sensitivity over higher energy neutrons so that patient dose may be significantly reduced.