1000

Asia Pacific J Clin Nutr (1995) 4: 187-189

Whole body measurement of C, N and O using 14 MeV neutrons and the associated particle time-of-flight technique

S. Mitra¹, J. E. Wolff¹, R. Garrett² and C. W. Peters³

1. Growth Physiology Group, AgResearch, Ruakura, Hamilton, New Zealand;
2. Department of Physics, University of Auckland, Auckland, New Zealand;
3. Nuclear Diagnostic Systems Inc., Springfield, VA, USA.

Our aim has been to construct a portable prototype instrument for measuring the whole body composition in vivo of growing lambs in terms of fat. protein and water by determining the mass of carbon, nitrogen and oxygen present. A small and compact sealed tube neutron generator which has the capability of exploiting the associated particle time-of-flight technique has been used for prompt gamma 14 MeV neutron activation analysis of C, N and O. This technique allows only gamma rays generated by neutron reactions within a defined volume to be recorded and offers a superior signal-to-noise ratio over existing prompt gamma neutron activation techniques. Based on the results obtained from irradiating a 41.4 kg meat phantom, we anticipate that an instrument comprising the neutron generator and four 15 x 15 x 45 cm Nal(TI) gamma ray detectors can be assembled to determine in vivo, protein, fat and water with precisions of 4.1, 5.4 and 1.2% (CV), respectively, within a 15 min scan. The radiation dose delivered would be ~0.03 mSv.

Introduction

Given the fixed stoichiometric proportions of N, C and O in protein, fat and water, proximate composition of the body can be derived from its elemental analysis. Despite use of in vivo neutron activation analysis for more than two decades to measure elemental composition of the human body^1-3, there is not yet an instrument which can measure N, C, and O during a single irradiation and detection procedure.

Instruments constructed during the 1980s have mainly used prompt gamma emissions to detect either N from the 10.83 MeV radiative capture emission¹ or in a few instances, C from the inelastic scattering reactions with 14 MeV neutrons^4,5. Kehayias et al.⁵ proposed that O be measured simultaneously. The major problems encountered have been the interferences and poor signal-to-noise ratios that result from scattered neutrons, and relatively high radiation doses.

We⁶ and Hollas et al.⁷ demonstrated the feasibility of using a coincidence system between the associated alpha particle and the 14 MeV neutron induced inelastic gamma ray to improve the signal-to-noise ratio so that N, C and O can be determined in large biological samples with good precision. Briefly, the alpha 1000 particle associated with a 14 MeV neutron produced by the ³H(d,n)⁴He reaction in the neutron generator is emitted in the opposite direction. Its detection specifies the neutrons in a given solid angle and, by appropriately gating the gamma detector, the resulting gamma ray spectrum is derived predominantly from neutron reactions within the defined volume of the sample.

A compact associated particle, sealed tube, neutron generator (APSTNG) with an internal alpha detector is now commercially available. Its success in detecting contraband drugs or explosives and use in coal analysis⁸ (where the elements of interest are similar to those for body composition) has been demonstrated. Here we report first results for measuring the C, N and O content of a large meat sample.

Methods

The APSTNG

The APSTNG and its high voltage control system was supplied by Nuclear Diagnostic Systems Inc., USA. It contained a mixture of deuterium and tritium, that was ionized by a Penning ion source, accelerated by a potential of 95 kV between the ion source cathode and target and focused to a spot (~1 mm diameter) on the target to produce neutrons and alpha particles by the ³H(d,n)⁴He reaction. Ion beam current (typically, 1 m A to obtain 10⁶ n/s) was controlled by heating a getter into which the ²H and ³H were absorbed. The tube also housed an alpha particle detector (3.7 cm ø ZnS screen positioned 4.5 cm from the target) which was interfaced to a Hamamatsu R580-15 fast rise time photomultiplier.

Data acquisition

Anode signals from both the alpha detector and a 12.5 cm diameter x 10 cm Bicron Nal(TI) gamma detector were each fed to timing filter amplifiers and their outputs to constant fraction discriminators. For alpha pulses, the discriminator was set just above the photomultiplier dark current with the APSTNG off while for gammas, the discriminator was set to reject 0.5 MeV pulses and below. Resulting logic pulses from alpha and gamma detectors, respectively, started and stopped the time-to-amplitude converter (TAC). Its output, the timeof-flight spectrum, was recorded with an analog to digital converter (ADC) and multichannel analyzer (MCA). The full width half maximum portion, corresponding to the time when most prompt gamma rays were emitted from a sample within the neutron beam, was selected by a single channel analyzer (SCA). The SCA output triggered a linear gate which allowed only gamma-ray energies within the defined time window to be recorded by an MCA and stored in a computer for subsequent spectral analysis.

Phantom construction

A 38.2 kg calibration phantom containing physiological amounts of the major body elements (ICRP 1975)⁹ was constructed. Its elemental composition by weight was 8.5% H, 18.2% C, 68.1% O, 2.7% N, 1.08% Ca, 1.02% P, 0.19% S, 0.17% K, 0.06% Na and 0.1% Cl. The solution was contained in polyethylene tubes (5 cm diameter x 80 cm) with a wall thickness of 0.15 mm and heat sealed. These were bundled into a cylindrical phantom (25 cm diameter by 80 cm) and wrapped with polyethylene sheeting. For a meat phantom of 41.4 kg we filled the same type of tubes with minced meat.

Irradiation and counting

The calibration and meat phantoms were each irradiated for 1 hour in scanning mode at a distance of 30 cm f 1000 rom the APSTNG. At this distance the neutron 'beam', defined by the associated alpha particle, was also 30 cm in diameter. The detector was positioned above the phantom and outside the defined neutron 'beam' with its face being 29 cm from the beam axis and perpendicular to it. The detector was shielded from direct neutrons by housing it in a 5 cm thick lead cave with an outer, 15 cm thick, layer of wax and borated polystyrene.

The total integrated counts in a region of interest (ROI) for the 1 hour irradiation were normalized to a constant neutron flux based on the recorded number of alpha counts. From this information, and assuming that sections of sample outside of the defined beam were shielded, the radiation dose during a 1 hour scan was calculated to be 0.12 mSv¹⁰.

Spectral analysis

Prompt gamma rays of interest from the fast neutron reactions: C; 4.43 MeV, N; 7.03, 5.1, 5.03 and 4.46 MeV and O; 7.12, 6.92, 6.13 and 4.43 MeV. Other intense gamma rays below 4 MeV are also detected but these suffer interferences from Ca, Cl, K and S (see Fig. 1). Consequently for the present report we have used only counts from the 4-7.5 MeV regions where the only elemental interferences are between C, N and O. Following the algorithm developed by Peters¹¹, ROI₁ was set at 5.35-6.25 MeV, ROI₂ was 4.85-7.28 MeV and ROI₃ was 4.22-4.68 MeV. Using liquid N₂ and H₂O we recorded pure spectra due to N and O, respectively, to define the interference coefficients. Background counts due to random coincidence events and from neutrons scattered from the sample into the detector within the defined time window were measured as follows: for ROI₁ and ROI₂ background counts were obtained from wax scatterers as there are no gamma rays from C above 4.68 MeV while for ROI₃, net counts were obtained by peak stripping from the underlying continuum.

Figure 1. The g spectrum of the meat phantom obtained within a 25 ns time window.

Results and discussion

In Table 1 we present estimates for radiation dose and the precision for measuring N, C and O in the meat phantom when compared with other detector geometries. Estimates for the four detector system for N utilizing the (n,g ) reaction are derived by scanning the same meat phantom at the Auckland Hospital, Body Composition Facility which uses two ²³⁸Pu/Be neutron sources and four 12.5 cm diameter by 15 cm Nal(TI) gamma ray detectors¹². The six detector system (see Table 1, third line) scales our results, obtained using the APSTNG and a single detector, to the same detector volume and scanning time as used at Auckland Hospital. The results show that precision for estimating N is comparable while radiation dose to the subject would be substantially reduced.

Table 1. Dose and precision of measuring N, O and C in a 41.4 kg phantom of minced meat with different NaI detector configurations.

			Precision, CV%, (n=10)
Number of 1000 detectors	Irradiation time	Dose msv	N	O	C
ONE (12.5cm ø x 10cm)	1 hour	0.12	11.5	2.3	9.2
FOUR (Auckland Hospital 12.5cm ø x 15cm)	30 min	0.2	5.0	--	--
SIX	30 min	0.06	6.7	1.4	5.4
FOUR (15 x 15 x 45cm)	15 min	0.03	4.1	<1	3.3

In order to obtain an acceptable precision for the measurement of body composition in farm animals, however, further enhancements of detection efficiency are required. In the fourth line of the table we have scaled our results from a one detector system to those which can be expected from the use of four large (15 cm x 15 cm x 45 cm) NaI detectors. These detectors are becoming available commercially and are obviously needed before the APSTNG system can achieve its full potential.

In Table 2 we compare analyses of meat composition by use of the APSTNG with existing procedures for proximate composition. The C, N and O content of the meat phantom was derived from the ratios of its elemental count rates with those obtained from the calibration phantom. Mass of protein was 6.25 x N while fat and water were determined from a four compartment model of body composition comprising protein, fat, water and minerals and their fractional content of C, N and O⁹. For this model, the small glycogen compartment has been ignored while 5% of phantom weight has been assumed for the mineral compartment⁹. Propagation errors that a 1000 rise from errors in the counting of each element are shown in Table 2. It can be seen that there is excellent agreement between the two methods of analysis for fat and water 2 while we have a significant discrepancy for protein. Reasons for this are being examined.

Table 2. Measured content of protein, fat and water in the meat sample.

Component	NAA (14MeV) kg	%	Chemical analysis %
Protein	6.0± 0.3 (4.1 %)	14.5± 0.6	16.5± 0.4
Fat	7.84± 0.42 (5.4%)	18.95+1.02	19.0± 0.8
Water	24.66± 0.29 (1.2%)	59.6± 0.7	59.7± 0.9
Sum		93.05	95.2
Ash			4.55± 0.57

Conclusion

A transportable instrument based upon the APSTNG and four 15 x 15 x 45 cm NaI(TI) detectors can be assembled for measuring the body composition of humans or farm animals in vivo. It would provide a simultaneous measure of protein fat and water to good precision within a 15 minute scanning time. The radiation dose (~0.03 mSv) would be almost one-tenth of the dose from existing facilities.

Acknowledgments—This work has been financially sup ported by the NZ Foundation for Research Science and Technology. Supply of the APSTNG and its High Voltage Control d56 System by Nuclear Diagnostic Systems (USA) at cost price and loan of the nuclear electronics by the Physics Department, University of Auckland is appreciated. We thank Professor G L. Hill and Dr L. D. Plank for use of the Auckland Hospital Body Composition facility.

References

1. Beddoe AH, Hill GL. Clinical measurement of Body Composition using in-vivo neutron activation analysis. J Parenteral Enteral Nutr 1985; 9: 504-520.
2. Cohn SH, Parr RM. Nuclear based techniques for the in-vivo study of human body composition. Clin Phys Physiol Meas 1985; 6: 275-301.
3. Chettle DR, Fremlin JH. Techniques of in-vivo neutron activation analysis. Phys Med Biol 1984; 29: 1011-1043.
4. Kyere K, Oldroyd B, Oxby CB, Burkinshaw L, Ellis RE, Hill GL The feasibility of measuring total body carbon by counting neutron inelastic scatter gamma rays. Phys Med Biol 1982; 27: 805-817.
5. Kehayias JJ, Zhuang H. Use of the Zetatron D-T neutron generator for the simultaneous measurement of carbon, oxygen and hydrogen in vivo in humans. Nucl Instr and Methods 1993; B79: 555-559.
6. Garrett R, Mitra S. A feasibility study of in vivo 14-MeV neutron activation analysis using the associated particle technique. Med Phys 1991; 18: 916-920.
7. Hollas CL, Ussery LE, Butterfield KB, Morgado RE. A method for in vivo determination of carbon and oxygen using prompt gamma radiations induced by 14.7 MeV neutrons. In: Yasumura S et al., eds. Advances in in vivo Body Composition Studies. New York: Plenum,1990: 395-400.
8. Gordon CM, Peters CW. A fast-neutron Probe for Tomography and Bulk Analysis. Int J Radiat Appl Instrum Part A 1990; 41: 1111-1116.
9. ICRP. Report of the Task Group on Reference Man. ICRP Report 23. Oxford; Pergamon, 1975: 289-327.
10. Goussev NG. Relationship between Dose Equivalent (Absorbed Dose) and Fluence (Flux Density). In: Jaeger IRG, ed. Engineering Compendium on Radiation Shielding, New York: Springer-Verlag, 1968; 1: 12.
11. Peters CW. Unpublished work.
12. Sutcliffe JF, Mitra S, Hill GL. In vivo measurement of total body carbon using 238Pu/Be neutron sources. Phys Med Biol 1990;35: 1089-1098.

Copyright © 1996 [Asia Pacific Journal of Clinical Nutrition]. All rights reserved.
Revised: January 19, 1999 .

to the top

0