Abstract
Aluminum, as an abundant metal, has gained widespread use in human life, entering the body predominantly as an additive to various foods and drinking water. Other major sources of exposure to aluminum include medical, cosmetic, and occupational routes. As a common environmental toxin, with well-known roles in several medical conditions such as dialysis encephalopathy, aluminum is considered a potential candidate in the causality of Alzheimer’s disease. Aluminum mostly accumulates in the bone, which makes bone an indicator of the body burden of aluminum and an ideal organ as a proxy for the brain. Most of the techniques developed for measuring aluminum include bone biopsy, which requires invasive measures, causing inconvenience for the patients. There has been a considerable effort in developing non-invasive approaches, which allow for monitoring aluminum levels for medical and occupational purposes in larger populations. In vivo neutron activation analysis, a method based on nuclear activation of isotopes of elements in the body and their subsequent detection, has proven to be an invaluable tool for this purpose. There are definite challenges in developing in vivo non-invasive techniques capable of detecting low levels of aluminum in healthy individuals and aluminum-exposed populations. The following review examines the method of in vivo neutron activation analysis in the context of aluminum measurement in humans focusing on different neutron sources, interference from other activation products, and the improvements made in minimum detectable limits and patient dose over the past few decades.
INTRODUCTION
Aluminum, being the most abundant metal and the third most common element in the earth’s crust, has a significant presence in our environment. Due to aluminum’s reactivity, it can only exist in bonded states in nature, most of which have very low water solubility, rendering low availability in surface waters. It can be found in a wide range of plants, drinking water, and food items, both natural and processed. The concentration of aluminum in processed food, however, is much higher because aluminum compounds are used as food additives, resulting in exposures that are more pronounced in industrialized societies. As a non-food, it has found uses in different industries and medical applications such as in antacids, aspirin coating, and vaccines, to name a few [1].
The routes of exposure for aluminum include ingestion, skin uptake, and inhalation, with the first being considered as the most common path of intake for the general public and the last being more significant in occupational settings. Its bioavailability, however, is quite low, as the gastrointestinal tract, the skin, and the lungs limit the absorption of aluminum to about 0.1% [1, 2]. Unlike its ubiquitous distribution in the environment, once absorbed, aluminum accumulates mostly in bone, lungs, muscle, liver, and brain, making the bone a favorable organ for measurement. Aluminum is categorized as a non-essential element for the human body; and excessive amounts of it can be toxic [3]. Aluminum neurotoxicity was known as early as 1942 through the work of Kopeloff [2, 4].
In the 1970s, there was a clear association between excess amounts of aluminum in the body and dialysis encephalopathy [2, 5]. Research conducted by various groups using different methods showed elevated aluminum concentrations in bones, sera, and brains of patients with renal failure who were undergoing dialysis therapy [5–8], some of whom had accumulated about 10 times more aluminum than healthy individuals [9]. In these patients, aluminum-contaminated water was used for preparation of the dialysate and aluminum-based phosphate binders were used in the treatment of hyperphosphatemia following dialysis [2, 10]. In the early publications, a death rate of 35–50% was reported in dialysis centers where large amounts of aluminum were used [2, 11]. Osteomalacia, characterized by bone pain and fractures, is another medical condition linked to dialysis. Osteomalacia seems to heal upon aluminum removal [6]. This has led to the investigation of other sources of exposure to aluminum and their possible effects.
Early studies based on animal models, as well as isolated cases of individuals exposed to aluminum, support a role for aluminum in Alzheimer-type dementia [12–14]. Twin studies indicate that the etiology of Alzheimer’s disease (AD) has both genetic and environmental components [15–18]. Because of its high prevalence, the sporadic or late-onset of AD has been largely associated with environmental toxins and exposures, and has made a stronger case for aluminum causality [19]. There is a clear need for further investigation of aluminum in the context of AD through development of accurate and sensitive techniques of measurement and clinical studies.
The link between aluminum and encephalopathy in dialysis patients became a motivation for investigation of aluminum in industrial and occupational exposures. One of the earliest studies, carried out in Sweden, investigated the concentration of aluminum in blood and urine samples collected from welders and aluminum production workers [20]. This showed higher concentrations of aluminum both in urine and blood samples compared to the healthy individuals that were nevertheless lower than the levels found in dialysis encephalopathy patients.
Another area of interest for measuring aluminum concerns long-term parenteral nutrition patients, for both infants and adults, where aluminum contamination of parenteral solutions poses the risk of aluminum toxicity and bone disease similar to that of dialysis patients [21]. Research findings in infants and adults receiving parenteral nutrition indicated elevated bone aluminum levels [22–24]. Controlling aluminum levels in patients receiving long-term parenteral nutrition and lowering their aluminum burden is of great importance, which requires routine monitoring. A non-invasive approach to aluminum monitoring is an invaluable tool in such cases, as bone biopsies cannot be considered as a method of routine monitoring due to the inconvenience to the patients.
Non-invasive aluminum monitoring could also prove to be very helpful in providing a screening technique for identifying patients with possible AD at an early stage. There is a clear need for development of methods that allow for accurate, sensitive, and reproducible measurement of aluminum in vivo. Currently, neutron activation analysis is the only available in vivo method for measuring aluminum concentration in human subjects.
With the growing interest in aluminum and its effects in the context of health and safety, different methods have been developed to measure aluminum accurately and with high sensitivity. Most of the studies on aluminum measurement use different methods of atomic absorption spectrometry [14, 26], mass spectrometry [27, 28], histochemical techniques [29], and instrumental neutron activation analysis [30–32]. All of these methods rely on in vitro approaches, and therefore, cannot be used on living humans.
An important consideration in measuring aluminum in vivo is determining the site of measurement, which depends strongly on the medical condition being investigated. Dialysis encephalopathy and AD are both conditions that are associated with elevated brain aluminum levels, subsequently making the brain an ideal site for measurement. In vivo measurement of aluminum in the brain of living subjects is not feasible due to the high doses of radiation required to penetrate the skull and the critical nature and radiosensitivity of the brain and its surrounding structures. Also, the measurement would have to be made by directing the neutron beam through the skull, which would generate an aluminum signal that would mask that from brain. As such, other sites need to be considered as a proxy for brain, with biokinetic and tissue deposition behavior of aluminum, feasibility of measurement, and ethical concerns taken into consideration. Aluminum can be measured in blood, more specifically in plasma, where it binds to transferrin, a protein associated with iron transport [1]. Urine is another medium for monitoring aluminum; however, it better reflects recent exposures and not the aluminum body burden [1]. Carrying about 60% of the aluminum body burden, the skeleton is considered the aluminum depot in the body, followed by the lungs, muscle, and liver [1]. The skeletal distribution provides a unique advantage in terms of the flexibility in the choice of measurement site. Additionally, AD has been associated with a higher risk of bone fractures, rendering the skeleton as an important site for measurement [33, 34].
Each of the above mentioned proxy systems present certain advantages, challenges, and limitations. In vitro systems such as urine and plasma come with the inherent risk of contamination due to the ubiquitous nature of aluminum in the environment, which begs the question of reliability and reproducibility, especially when low levels of aluminum are involved. Strict measures in sample collection and processing need to be taken in order to avoid contamination if the results are to be reliable. Most of the studies rely on a single sample collection, in order to reduce the inconvenience to the subjects, and therefore, reproducibility is most likely to be achieved through inter-laboratory collaborations. In vivo neutron activation measurement requires exposing subjects to ionizing radiation; this poses a limitation on dose and site of measurement. Extremities are considered the most suitable site due to their distance from critical and radiosensitive structures, as well as their higher regulatory dose limit [35]. Hand is a more convenient choice from the point of view of irradiation and shielding design, which plays an important role in the experimental setup. The major difference between the hand and the foot is that the hand is mainly comprised of cortical bone, whereas the foot, and especially the calcaneus, is trabecular bone. While aluminum accumulates in both bone types, higher concentrations of aluminum have been measured in trabecular, compared to cortical bone [36].
Bone biopsy for the purpose of measuring aluminum in living subjects, being an invasive method, is usually secondary to a surgical procedure. Therefore, the type of bone sample depends on the procedure and is often trabecular bone [37, 38]. Thus, in vivo measurement of aluminum in the trabecular bone would allow for inter-laboratory comparison of the results.
Defining typical values for bone aluminum based on the literature is challenging, as a wide range of concentrations measured using various methods have been reported by different researchers. To complicate matters further, based on the method used, the results are reported per gram of wet weight, dry weight, or relative to the calcium content of the bone, with no direct and accurate conversion factor available. Perhaps, the most comprehensive and relatively recent study conducted on measuring aluminum in bone is that of Hellstrom et al. [38]. Bone biopsy samples from 172 subjects suffering from different types of dementia including AD, as well as a control group were analyzed. Part of their results is shown in Table 1.
Also shown in Table 1 are aluminum levels in blood, urine, and brain samples in healthy individuals based on recent studies. There are considerable variations in values reported in the literature and between different subjects, as reflected by the wide range of aluminum content reported. In the case of the brain, varying concentrations of aluminum have been reported for different regions of the brain [12, 39]. As such, it is important to note that Table 1 is meant merely as a guide to provide the readers with estimates of aluminum content and the minimum levels that are likely to be detected in healthy humans.
IN VIVO NEUTRON ACTIVATION ANALYSIS
In vivo neutron activation analysis (IVNAA) is a method of measurement and analysis of living humans that can provide invaluable information for advances in medical research and well-being of mankind. The non-destructive nature of this method enables its application to humans. Additionally, it offers simultaneous measurement of several elements. It is based on the detection of characteristic gamma rays emitted during de-excitation of neutron-activated nuclei. It is limited, however, by the number of elements, which not only can become radioactive when bombarded by neutrons, but also have strong enough gamma emissions that can be detected using the current detection systems. For elements satisfying such criteria, IVNAA can become the method of choice, provided that measurements can be performed with sensitivity and accuracy that isdictated by the current exposure levels of those elements.
The first requirement for IVNAA is the neutron source for activating the elements of interest. Neutrons can be produced using a variety of sources (nuclear reactors, accelerators, and neutron-emitting radioisotopes) and reactions for the purpose of in vivo measurements, most of which are discussed in the review article by Chettle and Fremlin [41]. The activation rate, R, for a given element in a medium bombarded by neutrons is shown in the equation below:
Aluminum is one of the elements satisfying the criteria for IVNAA, by means of the nuclear reaction in which stable 27Al absorbs a neutron, forming radioactive 28Al. The relatively short half-life of this decay makes it a suitable radioisotope for in vivo measurement, as the detection time can be kept reasonably short. The relevance of a short half-life is that it allows for multiple of half-lives of the radioisotope to be acquired within a reasonable detection time, resulting in a more accurate analysis with smaller statistical uncertainty without requiring the participants to remain motionless for an extended period of time.
When irradiating a biological medium, in addition to the element of interest, other elements with significant neutron activation cross-sections will become activated. While this provides the advantage of measuring multiple elements simultaneously, some of the activation products may interfere with the radioisotope of interest. For the case of bone as the in vivo activation medium, the stable elements, activation reactions, and cross sections are listed in Table 2. Phosphorus, chlorine, and to a lesser extent sodium, are the major interfering reactions which affect the detection of aluminum in different ways. Phosphorus in particular presents a major challenge as 31P(n,α)28Al reaction shares the same activation product as aluminum, i.e., 28Al. Several issues must be considered in addressing the phosphorus interference. For the 31P(n,α)28Al reaction to take place, neutrons must have the minimum required energy dictated by the rest mass differences of the particles involved in the reaction giving a threshold energy of 2.01 MeV. Moreover, the probability of the alpha particle escaping the nucleus, also known as the Coulomb barrier, at 6.22 MeV for 31P(n,α)28Al, further inhibits the reaction. The total effect is such that there are no existing data showing a significant reaction rate for neutron energies below 3 MeV as shown in the excitation function (neutron absorption cross-section against neutron energy) in Fig. 1. It should be noted that P/Al activation ratio increases with increasing neutron energy and decreases with increasing sample size. The latter is due to the slowing down and moderation of neutrons in the body. The larger the amount of irradiated tissue, the more significant the moderation of neutrons will be. The phosphorus content in the bone and soft tissue is considerably higher than aluminum. According to ICRP 65, the reference man has 780 g and 0.06 g of phosphorus and aluminum, respectively [42]. Given the amount of phosphorous versus aluminum in the body, it is important preferably to eliminate the probability of reaction or devise other ways of accounting for its effect. 31P(n,α)28Al has a threshold of 2.01 MeV and as can be seen in the Fig. 1, there is no cross-section for this reaction for neutron energies below the threshold. This is important, as it provides a means for eliminating the interference, if neutrons of 2 MeV and above can be removed from the beam.
Another challenge when irradiating in vivo is the exposure of the subjects to ionizing radiation and the consequent dose. Precise measurement of the dose delivered to an individual is a complicated task in that both the dose from neutrons and the gamma dose from the neutron source or the target need to be taken into account.
The first attempt at using IVNAA goes back to 1964 and the work of Anderson et al. who measured Ca, Na, and Cl in two healthy individuals using 14 MeV neutrons [44]. They reported that while their Ca results agreed with the value reported for the ICRP’s (1959) “reference man”, the values for Cl and Na may have been overestimated. Their work was further expanded for other elements by Boddy et al. and other groups [45–48], but it was not until 1980 that IVNAA was used for measurement of aluminum.
East Kilbride group (14MeV neutron generator)
IVNAA has been used to measure whole body aluminum [49]. The system was previously developed and evaluated for measuring calcium, phosphorus, sodium, and nitrogen [46, 50]. Williams and the East Kilbridge group’s aim was to measure aluminum in subjects with chronic renal failure and healthy individuals. Their study cohort consisted of 38 subjects divided into three groups; subjects undergoing regular dialysis with and without dialysis encephalopathy and a control group not undergoing dialysis.
The criteria for dialysis encephalopathy included but were not limited to serum aluminum levels higher than 350μg/l. Their irradiation setup consisted of two 14 MeV neutron generator tubes placed above and below the patient. A whole body counter made up of two 29 cm dia×10 cm NaI(Tl) detectors was then used to detect the activity. The total dose delivered was 10 mSv at the body surface. As mentioned, the major drawback of using high energy neutrons in IVNAA is that the fast neutrons in the beam activate phosphorus which is a highly abundant element in the human body. However, measurements done on anthropomorphic phantoms revealed that 1 g of aluminum had the same signal as 8 g of phosphorus.
Irradiating the whole body provides significant moderation, thus partially thermalizing neutrons and giving rise to a low P/Al ratio. The remaining proportion of fast neutrons has a large enough contribution to phosphorus activation that aluminum can only be measured indirectly when 14 MeV neutrons are employed. The indirect measurement of aluminum then relies on the assumed more or less constant ratio of bone phosphorus to calcium (P/Ca) in healthy humans. Bone in adults contains almost the entire body burden of calcium and about 80% of phosphorus, in a ratio of 0.46±0.01 [51]. The remaining 20% is considered as non-bone phosphorus, in which any change is attributable to aluminum activation.
To calculate the non-bone phosphorus, the measured calcium multiplied by the P/Ca ratio (0.46) is subtracted from the total measured phosphorus. In order to account for variations in body size, the results are normalized to weight. Fluid retention and lean body mass are accounted for by normalizing to total body nitrogen and potassium, respectively.
One concern when using the standard P/Ca is to apply the ratio in healthy humans to cases of renal failure that commonly have bone disease and calcium depletion that in turn affects the P/Ca ratio. The error associated with such variations is reported as the ratio of measured to predicted calcium. The results showed significantly higher non-bone phosphorus in all subjects with chronic renal failure compared to the controls, while the serum phosphorus remained slightly above normal in the case of the former, which eliminated the possibility of increased phosphorus. There is also an insignificant increase in the non-bone phosphorus in subjects suffering from dialysis encephalopathy as compared to other dialysis patients.
Williams’ work set the first building block for in vivo measurement of aluminum and gave the first estimate of the upper limit for whole body aluminum of about 3 g. One advantage of the system was patient set-up. Having the subjects in the supine position provides comfort during measurement, in particular for measurements as long as 40 minutes. The other advantage of their system was its cost effectiveness. Neutron generators are a much cheaper alternative to using a reactor or a cyclotron, while offering relatively simpler operation with minimal warm up time [46].
While offering advantages in cost and operation, these 14 MeV neutron generators are not considered as a suitable source of neutrons due to the interference from phosphorus caused by the high energy neutrons and also their low neutron yield. Determining the aluminum content solely based on the non-bone phosphorus in the body may lead to an underestimation due to the fact that bone is the main depot for aluminum. Therefore, the contribution from the phosphorus in the skeleton cannot be ignored. Additionally, while the ratio of phosphorus to calcium in the bone is known to be constant in healthy individuals, it may be subject to considerable variation in the elderly, especially those suffering from bone ailments. The last, however, was accounted for by observing the ratio of measured to expected calcium in the subjects.
Brookhaven group (research reactor)
Ellis et al. improved the methodology for measuring aluminum in several ways [52]. Instead of using a neutron generator, they used a medical research reactor, which offered a thermal-epithermal neutron beam. The advantage of using a reactor beam is that it offers a higher neutron flux and, therefore, more activation and better detection. The beam used at Brookhaven had a thermal flux of about 107 n·cm–2·s–1. Using this beam allowed for activation of aluminum, though the presence of a minor component of fast neutrons in the beam still activated phosphorus. Moreover, they limited the radiation exposure to the subject’s hand, which reduced the dose to the patient while giving a reasonable estimate of aluminum burden in the body. The beam area was shielded down to about 10 cm×20 cm to accommodate a human hand. Following an irradiation time of 2 minutes and 2 minutes of delay for transferring to the counting facility, the activity in subjects’ hand was measured for 200 seconds. The detection system was another area of improvement where it used 4 NaI(Tl) detectors (4”×4”×16”) instead of two detectors in a quasi-4π geometry placed inside lead shielding. The minimum detection limit of the system was found to be 0.4 mg of aluminum based on phantoms containing relevant amounts of sodium, calcium and chloride compared to that of the “reference man”.
This study involved two separate groups of participants. The first group underwent total body neutron activation, and their whole body aluminum content was calculated based on total phosphorus and the fixed phosphorus to calcium ratio in the skeleton. The second group consisted of 10 subjects with total body measurements for calcium and phosphorus and partial body measurement for hand aluminum. Serum aluminum measurements were performed before and after administration of disferroxamine (DFX), which is an aluminum binder. These subjects had no clinical indication of aluminum-related diseases. Results of the hand irradiation were reported as the ratio of aluminum to calcium (Al/Ca) to account for variations in hand size. The aim of the total body measurement for this group was to provide an estimate of the total body calcium in order to convert from hand aluminum to total body aluminum. The serum Al ranges before and after DFX were 19–98μg/L and 32–450μg/L, respectively. The Al/Ca range for this group was 0.02–0.76 mg Al/g Ca, with a ratio of 0.04 or less considered within the normal range. For the same group, the total body Al ranged from undetectable to 2683 mg with a mean of 1350 mg. Skeletal Al burden range was 12–280 mg with a mean value of 162 mg. The mean total body and skeletal burden aluminum found in this group were higher than the reference man values [42] by factors of 8 and 20, respectively. The results of total body measurements done on 178 males showed a higher aluminum level based on total phosphorus for renal failure patients undergoing dialysis compared to the group suffering from renal disease but not on dialysis. They concluded that baseline and post-DFX serum aluminum levels did not correlate with total body aluminum, even though serum aluminum measurement is a common clinical method of estimating aluminum burden in the body. There was also no significant correlation between the total body aluminum and skeletal aluminum burden or Al/Ca ratio in the hand. They did, however, find a significant correlation between post-DFX serum aluminum and Al/Ca ratio in the hand. It is important to note that phosphorus contamination in the hand measurement was estimated to be equivalent to 0.7 mg of aluminum. For the total body irradiation, the anthropomorphic phantom study showed that 1 g of aluminum had the same signal as 68.3 g of phosphorus, due to the significant moderation of neutrons both in the shielding and within the body.
Birmingham group (accelerator)
Green and Chettle examined the feasibility of using an accelerator as a more available means of producing neutrons as compared to a reactor-based system,without a major compromise in the neutron flux [53]. They used the Dynamitron accelerator available at the University of Birmingham’s School of Physics and Space Research; a 3 MV machine that used a tritium target and its 3H(p,n)3He reaction to produce neutrons. The use of an accelerator offered the possibility of varying the incident proton energy and hence provided a neutron energy spectrum that enabled them to optimize neutron energy and, therefore, the dose.
The technique of microdosimetry allows for a detailed investigation of dose distribution in the irradiation cavity for various energies. It also has the advantage of providing information about the quality of radiation, also known as radiation weighting factor. Radiation weighting factor is a parameter that is used to convert the absorbed dose measured by radiation detection devices to equivalent dose. Equivalent dose is a quantity that is described to provide a common scale for biological damages of all types of ionizing radiation. This is because the extent of damage done by ionizing radiation can vary significantly with the type of radiation.
Two groups of phantoms with varying amounts of aluminum up to 100 mg or 130 mg were made with one group containing tissue-relevant concentrations of Na and Cl. The irradiation and counting protocol consisted of a 30-second irradiation time followed by transfer and counting times of 30 and 300 seconds, respectively. The detection system chosen was an assembly of four NaI(Tl) 51 mm dia×152 mm L detectors. Microdosimetry results showed that the optimum proton energy could be less than the lowest measured energy of 1.1 MeV. However, using lower energies was not practical due to cross-section restrictions of the target, causing a drastic drop in neutron fluence for energies less than 1.2 MeV. For this reason, all but one experiment (Ep = 1.05 MeV) were performed at the energy of 1.2 MeV. As expected, the NaI(Tl) detector assembly provided a better efficiency, and therefore, higher number of counts compared to germanium detectors.
The minimum detectable limit (MDL) achieved for 1.05 MeV and 1.2 MeV, were 2.0 mg and 1.2 mg, respectively. Comparing their MDL results with other published in vivo aluminum measurements, the authors concluded that using an accelerator with tritium target was a feasible method of measuring aluminum in vivo with suggested future work on improving the sensitivity of the system including detector upgrades, dose delivery and energy optimizations, and data analysis.
In a later study, Green et al. made improvements on dose measurement and also performed experiments comparing the phantom set made at BirminghamUniversity to that from Swansea group [54]. The Swansea phantoms were saline bags with physiologically relevant concentrations of Ca, Na, Cl, and P with varying amounts of Al (0–30 mg), whereas the Birmingham phantoms were smaller bottles of AlCl3 with aluminum ranging up to 130 mg. The detection system was changed to two NaI (Tl) detectors of different sizes, and the analysis was improved by employing a non-linear least-squares method. Two irradiation protocols of 60 seconds and 30 seconds corresponding to high and low dose delivery were used, followed by a 30-second transfer time and 300 seconds of counting time. The calibration lines generated from each set of phantoms revealed a difference in the slopes, which was attributed to the fact that the flexible containment of Swansea phantoms allowed for a closer positioning with respect to the detectors. Additionally, the measurements done on Swansea phantoms indicated that the interference from phosphorus was insignificant. The response of 1 g of phosphorus was equivalent to less than 0.1 g of aluminum and, therefore, was not considered as a concern for the irradiation protocol used.
Swansea group (252Cf neutron source)
In an attempt to make the technique of IVNAA more available, Wyatt et al. developed a system using 252Cf as the neutron source [55]. The main advantage in using 252Cf is that it is a readily available radionuclide that can be used as a source of neutrons, whereas reactors or accelerators are not widely available for clinical applications. The major disadvantages of 252Cf are the low thermal neutron yield and also the recurring issue of interference from the P reaction with fast neutrons. 252Cf has a neutron energy range of up to 10 MeV with an average energy of 2.2 MeV. Phantom measurements indicated that the interference from 7 g of P was equal to 1 g of Al. The mean neutron energy of 252Cf is lower than the14 MeV neutron generator, thus reducing phosphorus activation in comparison. However, partial-body irradiation of the hand means a smaller sample size and consequently, weaker moderation. In order to enhance the activation in favor of aluminum, the beam was moderated using water, which, while increasing the ratio of thermal to fast neutrons, reduced the thermal neutron flux. Water has moderating properties similar to that of the tissue, which provided the additional advantage of eliminating the need for tissue thickness correction. The detection system consisted of two NaI (Tl) detectors in a shielded enclosure with an opening for inserting the hand. System calibration using tissue equivalent phantoms with varying concentrations of aluminum was performed which verified the analysis technique.
Seven patients with a mean age of 54.4 years and range of 43–70 years were recruited for the clinical study, all of whom were suffering from end-stage renal failure and were either already on dialysis treatment or about to receive it. In an attempt to compensate for the low neutron yield, a cyclic technique was used. Four cycles of irradiation and counting of each 300 seconds were used, with a transfer time of 30 seconds between irradiation and counting. This technique improved the activation of aluminum by a factor of 2.5 compared to a single irradiation previously used by the same group and resulted in a dose equivalent of 36 mSv to the hand and an estimated total body dose equivalent of less than 0.2 mSv.
Wyatt et al. used the P/Ca ratios obtained by Ellis et al. [52] to account for the activation of phosphorus in the hand. Iliac bone samples from the same patients were analyzed for aluminum and calcium content using electrothermal atomic absorption spectrometry for comparison purposes. The range of aluminum found in the iliac bone samples was 47–108μg Al/g dry weight or 353–1121μg Al/g Ca. The results of the IVNAA were analyzed using library least squares fitting, based on the counts per gram for each of the elements in the tissue-equivalent phantoms, and were reported as the Al/ Ca ratio. The phantom results showed close agreement with the actual amounts of aluminum in each phantom. Al/Ca ratio in the hand ranged from –42 to 518μg Al/g Ca. The authors attributed the higher Al/Ca ratio in biopsy samples, compared to IVNAA of the hand, to the variations in the type of the bone measured. Iliac bone is trabecular, while hand is mostly cortical bone. The former has been shown to accumulate more aluminum but less calcium than the latter [36, 56]. IVNAA hand aluminum levels were in good agreement with the results obtained by Ellis et al. [52]. The MDL achieved was 180μg Al/g Ca or approximately 2.2 mg of aluminum in the hand, which is higher than the amount found in the hand of the reference man by a factor of 7 [42]. Such detection limit renders the system unsuitable for measuring aluminum in the normal population. However, it can be used successfully in patients suffering from renal failure as their bone aluminum levels can reach up to 50 times the amount found in healthy individuals [7].
McMaster group (McMaster research reactor and KN accelerator)
Palerme et al. [57] investigated the feasibility of using different neutron sources available at McMaster University with the goal of establishing a facility for measuring aluminum in bone using IVNAA. These sources included a nuclear reactor and a KN accelerator. Two of the reactor beams were investigated; one with a degraded fission spectrum, but no thermalization of neutrons. The other has silicon and sapphire crystal filtration that removes fast neutrons from the beam, hence reducing the probability of phosphorus activation.
The ratio of fast to thermal neutron flux was taken as indicative of the extent of aluminum production from silicon and phosphorus and samples of all three elements in powder form were measured. The KN accelerator used a lithium target with proton energy of 2 MeV, eliminating the activation of silicon and phosphorus due to the insufficient energy of the resulting neutron spectrum. The unfiltered reactor beam showed a clear contribution from silicon and phosphorus while the filtered beam had a 3.5% ratio of fast to thermal flux. While the accelerator provided the best results in terms of interference-free activation of aluminum, practical difficulties forced the experiments to be carried out using the filtered beam of the reactor.
Resin-based hand phantoms with the geometry of a clenched fist were constructed and irradiated for 3 minutes, and the data were subsequently acquired using two 200 mm diameter×50 mm thick NaI(Tl) detectors (120 mm of separation) after a transfer time of 45 seconds. Data analysis consisted of fitting two Gaussians to the chlorine and aluminum peaks and a single Gaussian to the calcium peak. Results were reported as the ratio of aluminum to calcium. The minimum detectable limit was calculated to be 1.5 mg of Al or 0.1 mg Al/g Ca. The neutron flux of 4×107 n·cm–2·s–1 resulted in a dose of 4 mSv to the phantoms. While this work improved the detection limit of aluminum by about 70% , presence of high energy neutrons in the beam and the subsequent interference from phosphorous remained a problem; a problem that could be solved using the accelerator available on site.
Comsa et al. [58] mainly focused on improving the MDL through data acquisition and analysis techniques at the already existing neutron activation analysis facility at McMaster University. They employed the method of spectral decomposition in order to take advantage of the total number of counts in the spectra, as opposed to the traditional photo-peak analysis. Single-element phantoms containing elements with significant thermal neutron cross-sections were constructed and irradiated to make up a library of spectra.
An algorithm was written to sum the contributions from each element in the fitted spectrum. Additionally, in order to improve the detection of aluminum, anelectronic coincidence rejection system was developed to diminish the effect of Cl and Na in the spectra.
Al emits a single gamma, whereas both Cl and Na emit gamma rays in a cascade. By rejecting the events that are detected in both detectors within the resolving time, the effects of both Na and Cl are greatly reduced. This is particularly important in the case of Cl, one of the peaks of which (1.64 MeV) is in the vicinity of the Al peak (1.78 MeV), as this method partially removes the overlapping of Al and Cl peaks in favor of Al. These modifications reduced the MDL to 0.7 mg of Al while keeping the equivalent dose at 20 mSv. It was shown that using the spectral decomposition technique improved the MDL by a factor of 1.7 when compared to the traditional method of least squares fitting. While an improvement to the previous work done, the MDL of 0.7 mg was still above the levels expected in the hand of a healthy individual.
Further improvements on the accelerator-based NAA system at McMaster paved the way for the pilot in vivo studies. One area of improvement included the development of a 4π detection system consisting of 8 NaI(Tl) detectors, the details of which are reported elsewhere [59, 60]. The new detection system together with the previous analysis upgrades, i.e., spectral decomposition, led to an estimated reduced MDL of 0.24 mg Al [61]. Furthermore, the Van de Graff accelerator used in the previous studies was replaced with a high-current Tandetron accelerator that made shorter irradiation times possible. The newly designed and constructed irradiation cavity, shown in Fig. 2, was another area of improvement expected to improve the MDL by a factor of 1.2–1.25 [61].
McMaster group (Tandetron accelerator)
Davis et al. [62] and Aslam et al. [63] performed pilot in vivo studies measuring aluminum in the hand of healthy and occupationally exposed individuals using the upgraded NAA system. A phantom study was performed prior to the in vivo study. Phantoms were made based on the cortical bone composition of ICRP 23 Reference Man and consisted of Ca, Na, Cl, Mg, and varying amounts of Al in a cylindrical container [42]. The phantoms were irradiated for 3 minutes, followed by a transfer time of 105 seconds, and were counted for 10 minutes. The equivalent and effective doses were determined to be 17.6 mSv (using a quality factor of 13 for neutrons) and 14.4μSv, respectively.
The pilot study consisted of 20 male volunteers residing in southern Ontario with no history of exposure to aluminum for medical or occupational purposes with a mean age of 51.8±13.1. The second group consisted of 8 subjects with self-reported occupational exposure to aluminum and a mean age of 56.6±5.4. The irradiation cavity was designed to irradiate the fingers and palm of the subject and an adjustable water sleeve was fitted around the arm to protect the subject from unnecessary exposure. The same irradiation protocol was used for the phantoms. Results were reported as Al/Ca ratio.
The MDL values determined were 0.29 mg Al or 19.5μg Al/g Ca for phantoms, which is at the low end of the range of 20–27μg Al/g Ca expected in the hand of healthy human. The MDL for the human hand was determined to be 28.1μg Al/g Ca while the range of aluminum concentration in the hand of healthy subjects varied between –9.6±11.6μg Al/g Ca and 60.3±10.4μg Al/g Ca with a mean value of 27.1±16.1μg Al/g Ca. The negative concentration was reported as undetectable. For the occupationally exposed, aluminum concentrations ranged from 34.2±19.8μg Al/g Ca to 44.6±18.5μg Al/g Ca, with a mean value of 41.2±4.5μg Al/g Ca. The difference between the two mean values was found to be significant at the 5% level.
The results obtained by Davis et al. [62] were inagreement with the aluminum levels in healthy individuals measured in vivo or in vitro by other groups. However, most of the measurements were very close to the detection limit of the technique which indicated that while their technique was capable of successfully measuring the aluminum concentration in subjects undergoing dialysis, or those exposed to high levels of aluminum, it was not as reliable in measuring healthy persons and control subjects. Further envisaged enhancements included reducing the irradiation and transfer time to improve the detection limit of the system.
There were two major improvements made to the system reported by Aslam et al. [63]. The first was the use of a high current accelerator, which allowed for reducing the irradiation time in order to increase activation and to minimize decay during irradiation, while maintaining the dose as low as possible. The current was increased from 100μA to 400μA and subsequently, the irradiation time was decreased to 45 seconds from the original 180 seconds. The second area of improvement involved the detection system shown in Fig. 3. The previous studies and also the first part of this study was done using an array of 8 NaI(Tl) detectors, two of which had a poor resolution which led to larger uncertainties in the measurements. These two detectors were replaced and an additional smaller detector was added to bring the solid angle even closer to 4π.
The phantom studies performed on the improved system also benefitted from a shorter transfer time of only 45 seconds. The importance of having a shorter transfer time lies in the fact that due to the 135-second half-life of aluminum, longer transfer times mean decay of the radionuclide during transfer which is undesirable from a detection standpoint. The total effect of increasing the current and reducing the irradiation and transfer times reduced the detection limit to 8.32μg Al/g Ca from the previously determined 19.5μg Al/g Ca, making the system capable of measuring much lower concentrations of aluminum likely to be found in unexposed individuals under current exposure levels.
FUTURE DIRECTIONS
The most recent improvements on the analysis technique and phantom measurements performed by Matysiak et al. at McMaster University were presented at the 10th Keele Meeting on Aluminum, held in Winchester, England in 2013. The main areas of improvement were: (1) using the technique of digital anticoincidence counting to reduce the interference from the chlorine peak by a factor of 4 [59], (2) using a half-life analysis, and (3) using low concentration phantoms which better resembled the aluminum levels found in human subjects.
Accelerator-based neutron beams offer many advantages by providing high fluence rates while eliminating or reducing the effect of interfering nuclear reactions. However, their limited availability in clinical settings, due to size and cost considerations, renders them impractical when large clinical studies or routine patient monitoring is of interest. Recently, Liu et al. [64, 65] investigated the feasibility of using a deuterium-deuterium (DD) neutron generator as a portable system for the in vivo measurement of manganese. The DD neutron source provides neutrons with the energy of about 2.45 MeV. Such a system can be very promising for IVNAA of aluminum, because of its technical characteristics, as well as cost and portability. Compared to a full-scale accelerator-based system, the portable DD neutron source can be purchased at a fraction of the cost of the standard-sized system without the need for a large space. The DD system has a high neutron yield compared to the 252Cf neutron source previously used by Wyatt et al. [55].
The neutron energy generated by the deuterium-deuterium fusion is also much lower than the 14 MeV neutrons used by Williams et al. [49] with the deuterium-tritium fusion reaction as the neutron source. While the 2.45 MeV neutrons generated by the DD system are above the energy threshold for the 31P(n,α)28Al reaction, considering the Coulomb barrier of 6.22 MeV for the reaction the probability of the activation of 31P would be low. The exact details of this activation reaction are currently being investigated at McMaster University. Given the above-mentioned advantages, the DD neutron generator has considerable potential for the measurement of aluminum.
Another area that requires further investigation is the inter-laboratory comparison of aluminum measurements. While IVNAA results are reported as the Al/Ca ratio, most of the chemical analysis techniques report the results in terms of mass ratio of aluminum to fresh, dry, ash, or wet bone. Parsons et al. [9] measured the ratio of ash/dry weight and dry/wet weight for control and renal failure, and dialysis samples; but to date, there are no validated conversion methods for comparing the results of different laboratories. The current method involves crude estimation of the amount of calcium in dry or wet bones, which may not lead to accurate results. Inter-laboratory collaborations in measuring the same samples with different methods and comparing the results can provide invaluable information about the aluminum body burden in different medical and occupational scenarios.
CONCLUSIONS
Due to the evidence for aluminum involvement in multiple medical conditions such as AD, as well as occupational hazards, more research needs to be done in the field of in vivo measurement of aluminum. More specifically, there is a need for clinical studies to test the new techniques and challenge their detection limits in clinical settings. For an IVNAA technique to be successful under the circumstances of a clinical study, certain technical criteria must be met. First and foremost, is the choice of a neutron source. An ideal neutron source has a high thermal neutron flux and zero or negligible high-energy neutrons to avoid the activation of 31P. This will maximize the activation of aluminum, while keeping the patient’s exposure to ionizing radiation as low as possible. While 252Cf enjoys the benefit of being more readily available, the current low levels of aluminum limit the choices to sources that offer a high flux of neutrons, such as thermalized reactor beams and accelerator beams, where the incident energy can be controlled. The choice of the site of measurement has important implications, both in terms of convenience and radiation safety. Bone is the ideal site for measurement of aluminum as most of the aluminum burden of the body is stored in the bone. The question of whether or not bone is a suitable proxy for brain needs to be addressed through postmortem analysis of both brain and bone samples of AD and control subjects, though this is outside the scope of in vivo aluminum measurement.
Ellis et al. is the only group comparing IVNAA with an alternative proxy system by reporting serum aluminum levels in renal failure patients as well bone aluminum [52]. Serum aluminum levels in healthy individuals are believed to be in the range of 1–3μg/L [40]. The baseline serum aluminum levels found in patients with renal failure in Ellis’ study were significantly raised compared to the healthy individuals; a result that was observed by Harrington et al. [66]. However, there was no correlation between baseline serum aluminum and the total body and skeletal aluminum levels. There was however, a significant correlation between the skeletal burden of aluminum and serum aluminum after administration of DFX. The reason for the lack of correlation in baseline values might be the fact that serum is considered as the transport system for aluminum and, therefore, may not accurately reflect long-term exposures and aluminum body burden. In general, serum and urine are better indicators of recent aluminum exposures [1].
Early IVNAA studies used the method of total body irradiation exposing the entire body to neutrons and hence delivering unnecessary dose to sensitive organs. The human hand has been the site of choice by many researchers for a variety of reasons. Human hand bone accounts for 1.5% of the skeleton. In general, extremities are not particularly radiosensitive [35] which makes the hand a suitable site for irradiation. The effective dose is also low due to the relatively small tissue weighting factors of skin and bone. Moreover, the hand is also a convenient site in terms of instrumentation and constructing the radiation cavity in such way that would protect the rest of the body from the harmful effects of radiation. The calcaneus bone in the foot is a less convenient choice which may be favored as trabecular bone has been shown to have a higher aluminum content compared to the cortical bone in the hand [36]. There is still need for further optimization and improvement in the field of IVNAA to make routine monitoring of bone aluminum possible.
Table 3 lists the progress of IVNAA studies in terms of MDL and the dose delivered to the subjects. As can be seen, even though the reactor provided a better neutron flux and lower MDL in the work of Ellis et al. in 1988 [52], the progress made over the next years on accelerator setup in terms of irradiation and detection geometries and analysis methods helped it to achieve even better detection limits and a lower dose to the subjects.
Table 4 summarizes the results of the clinical studies performed with IVNAA. While the continuous improvements in the detection limit over the years and its successful application in measuring bone aluminum in dialysis subjects is quite encouraging, a successful IVNAA system must be capable of detecting aluminum levels found in healthy subjects. Assuming the 1.06μg/g dry weight listed in Table 1 is the mean bone aluminum content of healthy individuals, a crude conversion factor of 4 can be adopted between the dry weight and calcium content of the bone. This is based on the ratio of the percentage of the inorganic component of the bone crystal that makes up the dry weight of the bone and the calcium content. Currently, the lowest MDL reported in the literature is 8.32μg Al/g Ca, reported by Aslam et al. [63]. An approximate aluminum content of 4.24μg Al/g Ca can be adopted for healthy individuals using the above conversion factor, which is lower than the current MDL by a factor of two. This is also indicated by the wide range of aluminum in healthy subjects as reported by Davis et al. [62] and shown in Table 4. These results further strengthen the need for achieving yet lower detection limits and conducting extensive clinical studies.
DISCLOSURE STATEMENT
Authors’ disclosures available online (http://www.j-alz.com/manuscript-disclosures/15-0595r2).