A clinical instrument for multi-element in vivo analysis by prompt, delayed and cyclic neutron activation using 252Cf

Optimisation of the body composition analyser using the gamma shield and improvement of accuracy using ORNL anthropomorphic phantoms

The analysis of body elements using the prompt gamma rays neutron activation method is a very useful and highly accurate method Lone, M. A., Mughabghab, S. F. and Paviotti-Corcuera, R. Development of a database for prompt gamma-ray neutron activation analysis. Summary Report of the Second Research Coordination Meeting. IAEA Headquarters. IAEA Nuclear Data Section. INDC(NDS)-424. (2001) that has many applications in different fields such as the diagnosis of specific diseases, including certain types of cancers. To protect patients from gamma rays produced by the system, an efficient strategy is to apply a gamma shield. In this study, the gamma shield was placed in three separate positions. The influence of these positions on reducing the effective dose was examined in a 5-y-old Oak Ridge National Laboratory mathematical phantom. Other parameters considered were sensitivity and coefficient of variation (CV) of thermal neutron fluence rate. With the best configuration, the total effective dose per minute (ET) was decreased ∼52 % and the sensitivity was ∼2.03-fold higher than when no shield was present.

Development of methods for body composition studies

This review is focused on experimental methods for determination of the composition of the human body, its organs and tissues. It summarizes the development and current status of fat determinations from body density, total body water determinations through the dilution technique, whole and partial body potassium measurements for body cell mass estimates, in vivo neutron activation analysis for body protein measurements, dual-energy absorptiometry (DEXA), computed tomography (CT) and magnetic resonance imaging (MRI, fMRI) and spectroscopy (MRS) for body composition studies on tissue and organ levels, as well as single- and multiple-frequency bioimpedance (BIA) and anthropometry as simple easily available methods. Methods for trace element analysis in vivo are also described. Using this wide range of measurement methods, together with gradually improved body composition models, it is now possible to quantify a number of body components and follow their changes in health and disease.

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.

Measurement of nitrogen in the body using a commercial PGNAA system—phantom experiments

An industrial prompt-gamma neutron activation analysis (PGNAA) system, originally designed for the real-time elemental analyses of bulk coal on a conveyor belt, has been studied to examine the feasibility of using such a system for body composition analysis. Experiments were conducted to measure nitrogen in a simple, tissue equivalent phantom comprised of 2.7 wt% of nitrogen. The neutron source for these experiments was 365 MBq (18.38 microg) of 252Cf located within an engineered low Z moderator and it yielded a dose rate in the measurement position of 3.91 mSv/h; data were collected using a 2780 cm(3) NaI(Tl) cylindrical detector with a digital signal processor and a 512 channel MCA. Source, moderator and detector geometries were unaltered from the system's standard configuration, where they have been optimized for considerations such as neutron thermalization, measurement sensitivity and uniformity, background radiation and external dose minimization. Based on net counts in the 10.8 MeV PGNAA nitrogen photopeak and its escape peaks the dose dependent nitrogen count rate was 11,600 counts/mSv with an uncertainty of 3.0% after 0.32 mSv (4.9 min), 2.0% after 0.74 mSv (11.4 min) and 1.0% after 3.02 mSv (46.4 min).

The use of associated particle timing based on the D + D reaction for imaging a solid object

Associated particle timing based on the D + D reaction has been applied for imaging a bulk sample, namely an aluminium box. The relatively low neutron energy, 2.8 MeV, allows a better spatial resolution from time-of-flight measurements. A combination of a Si detector for charged particles and an NaI(Tl) scintillator for inelastic-scatter gamma rays yielded an overall time resolution of 0.4 ns, giving a spatial resolution of better than 1 cm. A new reconstruction program was developed, yielding an image free from major artefacts.

The feasibility of accelerator-based in vivo neutron activation analysis of nitrogen

The feasibility of accelerator-based in vivo neutron activation analysis of nitrogen has been investigated. It was found that a moderated neutron flux from approximately 10 microA of 2.5 MeV protons on a 9Be target performed as well as, and possibly slightly better than the existing isotope-based approach in terms of net counts per unit subject dose. Such a system may be an attractive alternative to the widespread use of (238,239)Pu/Be or 252Cf neutron sources, since there is more flexibility in the energy spectrum generated by accelerator-based neutron sources. From a radiation safety standpoint, accelerators have the advantage in that they only produce radiation when in operation. Furthermore, an accelerator beam can be pulsed, to reduce background detected in the prompt-gamma measurement, and such a device has a wide range of additional biological and medical applications.

Of mermaids and mountains. Three decades of prompt activation in vivo

During 1966 to 1972, several laboratories demonstrated the feasibility of measuring the major body elements H, N, Ca, and Cl by prompt gamma in vivo neutron activation analysis (PGIVNA). The MERMAID facility in Birmingham, England used a cyclotron-produced pulsed neutron beam, but other groups in the United Kingdom, United States, Canada, and New Zealand subsequently developed systems based on radioisotope neutron sources that could measure body nitrogen with a precision of a margin of error of a few percentage points. The accuracy of N measurement was greatly enhanced by Vartsky's internal standardization, using prompt-gamma H as the marker and total body hydrogen (based on total body water and skinfolds) as the reference. Chlorine and extracellular water volume were used in a similar way by the Swansea group to calibrate the prompt-gamma analysis of total body calcium. The PGIVNA technique is most valuable in assessing nutritional status, particularly in relation to body protein.

Assessment of background hydrogen by the Monte Carlo computer code MCNP-4A during measurements of total body nitrogen

The use of a hydrogen internal standard to enable the estimation of absolute mass during measurement of total body nitrogen by in vivo neutron activation is an established technique. Central to the technique is a determination of the H prompt gamma ray counts arising from the subject. In practice, interference counts from other sources--e.g., neutron shielding--are included. This study reports use of the Monte Carlo computer code, MCNP-4A, to investigate the interference counts arising from shielding both with and without a phantom containing a urea solution. Over a range of phantom size (depth 5 to 30 cm, width 20 to 40 cm), the counts arising from shielding increased by between 4% and 32% compared with the counts without a phantom. For any given depth, the counts increased approximately linearly with width. For any given width, there was little increase for depths exceeding 15 centimeters. The shielding counts comprised between 15% and 26% of those arising from the urea phantom. These results, although specific to the Swansea apparatus, suggest that extraneous hydrogen counts can be considerable and depend strongly on the subject's size.

Human body composition: in vivo methods

In vivo methods used to study human body composition continue to be developed, along with more advanced reference models that utilize the information obtained with these technologies. Some methods are well established, with a strong physiological basis for their measurement, whereas others are much more indirect. This review has been structured from the methodological point of view to help the reader understand what can be examined with each technique. The associations between the various in vivo methods (densitometry, dilution, bioelectrical impedance and conductance, whole body counting, neutron activation, X-ray absorptiometry, computer tomography, and magnetic resonance imaging) and the five-level multicompartment model of body composition are described, along with the limitations and advantages of each method. This review also provides an overview of the present status of this field of research in human biology, including examples of reference body composition data for infants, children, adolescents, and adults.

The elemental analysis of the human body using associated particle timing based on the D + D reaction

Fast (approximately 2.8 MeV) neutrons are generated in the 2H(d,n)3He reaction, and are used for excitation of sample materials via inelastic scattering. Nanosecond timing identifies the spatial origin of the measured gamma radiation, making possible three-dimensional imaging and improved detection sensitivity. Several elements have been measured, and it is estimated that a clinical device, with equivalent dose of 10 mSv, will be capable of detecting the order of 10-100 mg.

Benchmarking the MCNP code for Monte Carlo modelling of an in vivo neutron activation analysis system

The Monte Carlo computer code MCNP (version 4A) has been used to develop a personal computer-based model of the Swansea in vivo neutron activation analysis (IVNAA) system. The model included specification of the neutron source (252Cf), collimators, reflectors and shielding. The MCNP model was 'benchmarked' against fast neutron and thermal neutron fluence data obtained experimentally from the IVNAA system. The Swansea system allows two irradiation geometries using 'short' and 'long' collimators, which provide alternative dose rates for IVNAA. The data presented here relate to the short collimator, although results of similar accuracy were obtained using the long collimator. The fast neutron fluence was measured in air at a series of depths inside the collimator. The measurements agreed with the MCNP simulation within the statistical uncertainty (5-10%) of the calculations. The thermal neutron fluence was measured and calculated inside the cuboidal water phantom. The depth of maximum thermal fluence was 3.2 cm (measured) and 3.0 cm (calculated). The width of the 50% thermal fluence level across the phantom at its mid-depth was found to be the same by both MCNP and experiment. This benchmarking exercise has given us a high degree of confidence in MCNP as a tool for the design of IVNAA systems.

Monte Carlo simulation of prompt gamma neutron activation analysis using MCNP code

Prompt gamma neutron activation analysis (PGNAA) is the most direct method of measuring total-body nitrogen. In combination with internal hydrogen standardisation, it is possible to reduce the dependence on body habitus. The uniformity of activation and detection, however, cannot be optimised sufficiently to eliminate the dependence entirely, and so further corrections are essential. The availability of the powerful Monte Carlo code MCNP(4A) has allowed a more accurate analysis of the activation facility, and yields corrections for body habitus and superficial fat layers. The accuracy of the correction is retained as the source-to-skin distance is reduced, although the activation uniformity is thereby degraded. This allows the use of a 252Cf source with lower activity and hence reduces the running cost of the facility.

Improvement of the prompt-gamma neutron activation facility at Brookhaven National Laboratory

The prompt-gamma neutron activation facility at Brookhaven National Laboratory was upgraded to improve both the precision and accuracy of its in vivo determinations of total body nitrogen. The upgrade, guided by Monte Carlo simulations, involved elongating and modifying the source collimator and its shielding, repositioning the system's two NaI(Tl) detectors, and improving the neutron and gamma shielding of these detectors. The new source collimator has a graphite reflector around the 238PuBe neutron source to enhance the low-energy region of the neutron spectrum incident on the patient. The gamma detectors have been relocated from positions close to the upward-emerging collimated neutron beam to positions close to and at the sides of the patient. These modifications substantially reduced spurious counts resulting from the capture of small-angle scattered neutrons in the NaI detectors. The pile-up background under the 10.8 MeV 14N(n, gamma)15N spectral peak has been reduced so that the nitrogen peak-to-background ratio has been increased by a factor of 2.8. The resulting reduction in the coefficient of variation of the total body nitrogen measurements from 3% to 2.2% has improved the statistical significance of the results possible for any given number of patient measurements. The new system also has a more uniform composite sensitivity.

Monte Carlo design study of a moderated 252Cf source for in vivo neutron activation analysis of aluminium

The Monte Carlo computer code MCNP has been used to design a moderated 252Cf neutron source for in vivo neutron activation analysis of aluminium (Al) in the bones of the hand. The clinical motivation is the need to monitor Al body burden in subjects with renal dysfunction, at risk of Al toxicity. The design involves the source positioned on the central axis at one end of a cylindrical deuterium oxide moderator. The moderator is surrounded by a graphite reflector, with the hand inserted at the end of the moderator opposing the source. For a 1 mg 252Cf source, 15 cm long x 20 cm radius moderator and 20 cm thick reflector, the estimated minimum detection limit is 0.5 mg Al for a 20 min irradiation, with an equivalent dose of 16.5 mSv to the hand. Increasing the moderator length and/or introducing a fast neutron filter (for example silicon) further reduces interference from fast-neutron-induced reactions on phosphorus in bone, at the expense of decreased fluence of the thermal neutrons which activate Al. Increased source strengths may be necessary to compensate for this decreased thermal fluence, to allow measurements to be made within an acceptable time limit for the comfort of the patient.

The use of associated particle timing based on the D + D reaction for elemental analysis of bulk samples such as the human body

The use of associated particle timing based on the D + D reaction has been demonstrated for elemental analysis of bulk samples such as the human body. The neutron energy of 2.8 MeV eliminates the background from organic matrices. The nanosecond timing of a HPGe detector renders it possible to identify the spatial origin of the measured gamma radiation limiting the sensitive area to a single pixel. By this technique the background could be reduced by a factor of > or = 1000, but the present set-up has achieved an effective factor only in the range 20-100, due to losses in the generation of timing signals. The very clean gamma-spectra obtained permit the use of high efficiency scintillation detectors. Sensitivities for measuring Al, Ti, and Fe are presented at an extrapolated dose of 10 mSv.

Absolute measurements of total-body calcium using prompt gamma neutron analysis and internal chlorine standardization: results in renal patients

Measurement of total-body calcium (TBCa) is important in the study of several medical conditions. Neutron activation analysis (NAA) is sufficiently precise for serial measurements, but in order to provide single, absolute determinations of TBCa, it requires a correction for body habitus. A method is described for the absolute determination of TBCa by comparing the total-body chlorine, obtained by NAA, with that obtained by stable bromine dilution. The method is applied to a group of thirty patients who had been treated for chronic renal failure, and significant differences are found between measured values of TBCa and those predicted for 'normal' healthy subjects. The overall experimental error in the absolute determination of TBCa is estimated to be 8%, for a dose equivalent of 3.1 mSv (based on skin dose, and a quality factor of 10 for neutrons).

The effect of oestradiol implants on regional and total bone mass: a three-year longitudinal study

OBJECTIVE

Although there is evidence from cross-sectional studies that percutaneous oestrogen administration protects against menopausal bone loss, few longitudinal data are available. We have examined the effect of 3 years' treatment with percutaneous oestradiol on total body calcium, spinal trabecular bone mineral density and radial bone mineral content in post-menopausal women.

DESIGN AND PATIENTS

Twenty-nine post-menopausal women, aged 37-55 years, who had undergone hysterectomy and had experienced the onset of menopausal symptoms within the previous 2 years, were studied before and for 3 years during hormone replacement with oestradiol implants, given at approximately 6-monthly intervals.

MEASUREMENTS

Total body calcium was measured by prompt gamma neutron activation analysis, spinal trabecular bone mineral density by quantitative computed tomography and radial bone mineral content by single-photon absorptiometry.

RESULTS

There was a significant increase in the mean total body calcium, spinal trabecular bone mineral density and radial bone mineral content over the 3 years of the study. The mean (+/- SEM) percentage change per annum was +2.4% (+/- 0.8) for total body calcium (P < 0.01), +3.3% (+/- 0.6) for spinal trabecular bone mineral density (P < 0.001) and +1.2% (+/- 0.6) for radial bone mineral content (P < 0.05).

CONCLUSIONS

Percutaneous oestradiol replacement therapy prevents menopausal bone loss and is associated with a sustained and significant increase in total body calcium, spinal trabecular bone mineral density and radial bone mineral content over a 3-year treatment period. Oestradiol implants thus have skeletal effects comparable to those of oral or transdermal oestrogens.

The development of a technique to measure bone aluminium content using neutron activation analysis

The accumulation and toxicity of aluminium in patients with chronic renal failure is a well recognized hazard, and there is a need for a non-invasive technique to assess Al tissue load in these patients. The technique of in vivo neutron activation analysis, using a thermal neutron beam from a reactor, has been employed by previous workers, who measured Al in the hand with a detection limit of 0.4 mg for a dose equivalent of 20 mSv. However, the application of this technique is restricted by the very limited availability of nuclear reactors. We report the modification of an existing 252Cf-based instrument and construction of a shielded, high-efficiency counting system for the in vivo measurement of Al in the hand. Phantoms containing tissue-equivalent solutions of Ca, P, Na and Cl with various Al loadings were used for validation of the technique. The Al/Ca ratio in the hands of seven patients with renal failure was measured using a cyclic activation technique to compensate for the relatively low neutron output of the 252Cf source, and a detection limit of approximately 2.2 mg Al was achieved for a dose equivalent of 36 mSv. The results were compared with the Al content of iliac crest bone biopsy specimens measured using electrothermal atomic absorption spectrophotometry.

Determination of whole-body nitrogen and radiation assessment using in vivo prompt gamma activation technique

Body nitrogen content in the phantom is measured by semiconducting and scintillation spectrometers using in vivo prompt gamma-ray activation analysis technique. The effective dose rate equivalents for sensitive organs and tissues inside the phantom are assessed by dosimetric measurement and neutron transport calculation. The bismuth germanate scintillator is found superior to the germanium semiconducting detector to quantitatively measure the photopeak of the 10.829 MeV prompt gamma-ray emitted from the 14N(n, gamma) reaction. Recommended scanning period for current setup using the BGO detector is 1 h on the modified mobile nuclear reactor. The effective dose equivalents from both neutrons and gamma-rays are estimated around 63 microSv per scan in the phantom test, making it a safe and reliable nuclear analytical method for in vivo body nitrogen measurement.

The measurement of silicon in a lung phantom--a comparison of two nuclear reactions for in vivo activation analysis

The amount of silica in the human lung may be estimated by measurement of silicon using in vivo neutron activation analysis. A pulsed, fast neutron beam, produced with a 2 MV Van de Graaff generator using the 2H + 2H reaction, was used to irradiate a Si-doped chest phantom in order to determine minimum detection limits (MDL). Two 'in-beam' nuclear reactions on Si were studied; prompt fast neutron inelastic scatter 28Si (n,n' gamma)28Si reaction was measured during the beam burst and the slow neutron prompt capture reaction was measured between the fast neutron bursts. Although the latter reaction appeared less favourable due to neutron cross section and measurement efficiency considerations, it yielded an MDL of 1.8 g compared with 2.3 g for the 28Si(n,n' gamma)28Si reaction. A comparison was made with a 252Cf neutron irradiation system where a Si MDL of 6.3 g was obtained using the slow neutron capture reaction. The Van de Graaff system permits 'exposed' Si lung burdens to be measured but not normal levels. Improved measurement sensitivity may be achieved by reduction of high counting-rate losses and high background radiation.

Calibration for measuring total body nitrogen with a newly upgraded prompt gamma neutron activation facility

A description is given of the calibration and performance of the upgraded facility at Brookhaven National Laboratory (BNL) for measuring total body nitrogen using the technique of prompt gamma neutron activation analysis. With the improved calibration, total body nitrogen can be more accurately measured not only in normal subjects but also in obese and wasted patients. Body hydrogen is used as an internal standard. We examined the effect of a heavy-water premoderator on the uniformity of composite sensitivity, nitrogen and hydrogen measurement statistics, and dose to the subject. The calibration technique corrects the ratio of nitrogen-to-hydrogen counts measured from the subject for body size. Additionally, a correction for subcutaneous adipose tissue on the nitrogen-to-hydrogen count ratio is introduced. The newly upgraded BNL facility provides precision in counting statistics using a Remcal anthropomorphic phantom filled with a tissue-equivalent solution of 2.1% for a body dose of 0.35 mSv using a 2.5 cm D2O premoderator. The measurements were made at five 20 cm sections, counting for 200 s per section.

Design studies related to an in vivo neutron activation analysis facility for measuring total body nitrogen

Design studies relating to an in vivo prompt capture neutron activation analysis facility measuring total body nitrogen are presented. The basis of the design is a beryllium-graphite neutron collimator and reflector configuration for (alpha, n) type radionuclide neutron sources (238PuBe or 241AmBe), so as to reflect leaking, or out-scattered, neutrons towards the subject. This improves the ratio of thermal neutron flux to dose and the spatial distribution of thermal flux achieved with these sources, whilst retaining their advantage of long half-lives as compared to 252Cf based systems. The common problem of high count-rate at the detector, and therefore high nitrogen region of interest background due to pile-up, is decreased by using a set of smaller (5.1 cm diameter x 10.2 cm long) NaI(Tl) detectors instead of large ones. The facility described presents a relative error of nitrogen measurement of 3.6% and a nitrogen to background ratio of 2.3 for 0.45 mSv skin dose (assuming ten 5.1 cm x 10.2 cm NaI(Tl) detectors).

A total body nitrogen facility for paediatric use

The design, calibration and evaluation of a facility for in vivo prompt gamma neutron capture analysis of total body nitrogen in children is described. The patient is scanned in both supine and prone positions, across a vertically collimated beam from a 1 GBq 252Cf fission source. Two NaI(T1) detectors are placed on either side of the patient, perpendicular to both the neutron beam and the scanning direction. The effective dose equivalent delivered to a child during an 840 s scan is approximately 0.14 mSv (QF = 10). Correction factors for nitrogen background (width-dependent), hydrogen background (1-4% of hydrogen gamma ray peak) and the differential attenuation of nitrogen and hydrogen gamma rays (width-dependent) can be applied to the measured nitrogen-to-hydrogen gamma ray counts ratio. By using the mass of hydrogen (based on body mass and fat mass) as an internal standard, the nitrogen mass can then be determined. Measurements with a urea-containing box phantom show that the current precision (CV) of the net nitrogen counts and of the nitrogen-to-hydrogen counts ratio is +/- 2.0% and +/- 1.5%, respectively. Using small anthropomorphic and other phantoms, estimation of the mass of nitrogen has a precision of +/- 1.4 to 5.4% and an accuracy of 97.1 to 101.5%.

An improved in vivo neutron activation system for measuring kidney cadmium

An in vivo neutron activation system for measuring kidney cadmium has been redesigned, firstly to reduce ambient dose levels and, secondly, to improve the cadmium signal to neutron dose ratio by modifying the neutron spectrum from 238Pu/Be, by interposing a beryllium premoderator. The ambient dose was reduced by a factor of seven. The overall system performance (lower limit of detection for a given dose) was improved by 40-50%. This 238Pu/Be based system now performs as well as or better than analogous 252Cf systems, without the drawback of the relatively short half life of 252Cf.

In vivo measurements of cadmium and lead in occupationally-exposed workers and an urban population

This paper reports the preliminary findings of a survey of lead and cadmium body burdens in a nonoccupationally exposed population in Swansea, Wales, using the techniques of in vivo neutron activation and X-ray fluorescence analysis. Some measurements on an occupationally cadmium-exposed group are also included. The results confirm the association between cadmium and smoking and bone lead and age. The in vivo measurements demonstrate a degree of comparability with other data, which supports the further detailed analysis of the relationships between body burden and exposure, on the one hand, and possible health effects on the other.

Measurements of total body calcium by prompt-gamma neutron activation analysis using a 252Cf source

A clinical neutron activation instrument has been developed for in vivo elemental analysis. Utilizing the prompt-capture gamma ray technique, simultaneous total body (TB) measurements of primarily Ca, but also Cl, N, C, and H are routinely performed. This paper describes a technique for the measurement of TBCa (g) that relies on the use of TBCl as an internal standard. The method has been tested with four anthropomorphic phantoms covering a range of body habitus. The mean discrepancy between the measured and known Ca contents was 3.6%. The technique has been applied to two patient groups, and encouraging results were obtained.

Determination of total body calcium by prompt gamma neutron activation analysis: absolute in vivo measurements

Sequential TBCa measurements using the prompt gamma ray technique are firmly established and a number of patient groups are under investigation. A technique has been proposed for the absolute measurement of TBCa based upon the use of TBCl as an internal standard. Further work is needed to establish the validity of the method, but the initial results are encouraging.

An introduction to body composition models used in nutritional studies

Chemical analysis of cadavers has led to the definition of 'reference' man but diseases may result in significant changes in body composition. By measuring body composition in patient groups the nature of disease progression can be followed and management regimens evaluated. Various techniques are available which attempt to measure body composition in vivo. Several models of body composition have evolved with the introduction of new measurement techniques. A description of these models is presented. The limitations of the models and the techniques adopted for their measurement are discussed.

Calibration and evaluation of a 252Cf-based neutron activation analysis instrument for the determination of nitrogen in vivo

A novel clinical instrument for multi-element in vivo neutron activation analysis has been recently constructed in Swansea. The instrument is intended primarily for prompt gamma measurement of total and partial body calcium, total body nitrogen and partial body cadmium. For the measurement of nitrogen the subject is scanned both prone and supine across a vertical collimated neutron beam from a 4 GBq 252Cf source. Two shielded Nal(TI) detectors, each of volume 2760 cm3, are placed above the subject on the opposite side to that irradiated. The prompt gamma ray spectrum contains prominent peaks from hydrogen, nitrogen, chlorine and carbon. The optimisation, calibration and evaluation of the instrument for the measurement of nitrogen, by the reaction 14N(n, gamma)15N, is described. The calibration corrects the ratio of nitrogen-to-hydrogen counts measured from the subject for background gamma rays and the effects of body habitus. Body hydrogen is use as as internal standard. Repeated measurements of a homogeneous anthropomorphic phantom indicate that the ratio of nitrogen-to-hydrogen counts may be determined by a coefficient of variation of 1.6% for a neutron dose equivalent incident on the phantom of 0.45 mSv (QF = 10). The accuracy of the calibration was assessed by measuring three anthropomorphic phantoms (weight range: 41.4-110 kg) containing simulated skeletons and the major organs of the body. For these phantoms the mean discrepancy of the measured to the known nitrogen content was +4.9%. The simultaneous measurement of chlorine and carbon is discussed.