Introduction

The term nuclear data includes all data which describe either the properties of nuclei or their interactions. In general, all those data can be grouped under three headings, namely, nuclear structure, nuclear decay and nuclear reaction data. Accurate nuclear data are needed for applications in energy-related research (fission, fusion, accelerated driven systems, etc.) as well as non-energy related studies (medical radionuclide development, radiation therapy, astrophysics, etc.). The scope of this article is limited to medical applications, the basic aim of which is to provide a fundamental base for optimum production and application of radionuclides. The activities consist of new experimental measurements, nuclear model calculations, and standardisation and evaluation of existing data [for earlier reviews cf. 1–5].

As regards radioactive decay data, continuous experimental and evaluation efforts are underway [for a recent review cf. 6], and good quality data are available, for example from Evaluated Nuclear Structure and Decay Data File (ENSDF), BNL, USA. Only occasionally in case of some emerging radionuclides small discrepancies are observed, e.g. in the intensities of weak γ-ray lines, β+ emission branching or Auger electron spectra [cf. 7]. Specific attempts then need to be undertaken to remove those discrepancies. It should be emphasized that the very choice of a radionuclide for medical application depends on its decay data.

With regard to nuclear reaction data for radionuclide production, mainly neutron and charged-particle induced reactions have been of interest. In recent years, however, some attention has also been paid to photon-induced reactions. The processes involved in neutron interactions are (n,γ) (n,f) and (n,p), and extensive data are available, for example from Evaluated Nuclear Data File (ENDF/B-VII.0), BNL, USA. For charged-particle induced reactions also the experimental data base has been considerably strengthened in recent years (cf. EXFOR, IAEA, Vienna). In contrast there is still paucity in photon-induced reaction data [cf. 8]. It should be pointed out that nuclear reaction data play a very important role in the optimisation of a production procedure, i.e. maximising the yield of the desired product and minimising the level of impurities [cf. 1, 2].

In this article the status of data of radionuclides used in nuclear medicine is briefly reviewed, with emphasis on data for research-oriented radionuclides.

Commonly used radionuclides in nuclear medicine

Status of nuclear data

Diagnostic radionuclides

The underlying principle in diagnostic nuclear medicine is that the radiation dose to the patient is as low as possible, compatible with the required quality of imaging and the diagnostic advantage in comparison to non-radioactive methods. To this end, in Single Photon Emission Computed Tomography (SPECT), γ-ray emitting radionuclides like 99mTc (T ½ = 6.0 h), 123I (T ½ = 13.2 h) and 201Tl (T ½ = 3.06 d) are commonly used. In positron emission tomography (PET), on the other hand, short-lived β+-emitting radionuclides like 11C (T ½ = 20.4 min), 13N (T ½ = 10.0 min), 15O (T ½ = 2.0 min), 18F (T ½ = 110 min), 68Ga (T ½ = 68.3 min) and 82Rb (T ½ = 1.3 min) often find application. The radionuclides 99mTc, 68Ga and 82Rb are available through generator systems. The parent of 99mTc (i.e. 99Mo) is produced using in a nuclear reactor but the parents of 68Ga and 82Rb (i.e. 68Ge and 82Sr) are produced using intermediate energy protons from a cyclotron or an accelerator. All of the other radionuclides are produced via direct charged-particle induced reactions at cyclotrons. The decay and production data of all those radionuclides are well known [cf. 2, 9–11].

As an example, the data for the 18O(p,n)18F reaction, which is now commonly used for the production of the most widely used PET radionuclide 18F, are shown in Fig. 1A. The reaction was investigated both by neutron detection and residual activity measurement. The neutron data show better resolution but are difficult to quantify. The resonances are due to population of several high-lying states in the product nucleus. First systematic activation measurements by Ruth and Wolf [12], followed by detailed studies by Hess et al. [13], showed resonances at 5.1, 6.1 and 7.2 MeV which are in agreement with those found in spectral measurements of neutrons. It should be pointed out that due to resonance structure nuclear model calculations cannot reproduce excitation functions of reactions on light nuclei with reliable accuracies, calling for good experimental studies. On the other hand, while calculating the integral yield of the product, the structures disappear and a smooth curve is obtained (see Fig. 1B). The extended data of Hess et al. [13] up to 30 MeV should allow construction of large targets to produce 18F in quantities up to 500 GBq (~14 Ci) or even more.

Fig. 1
figure 1

A Excitation function of the 18O(p,n)18F reaction. Results of both neutron and activation measurements are shown. The rather bold curve is an eye-guide to the activation data (taken from Ref. [13]). B Thick target saturation yield of 18F calculated from the data by Ruth and Wolf [12] and those by Hess et al. [13]

Full size image

Therapeutic radionuclides

The spectrum of radionuclides commonly used in internal radionuclide therapy (endotherapy) is very broad. Since in this case a localised, well-defined radiation dose needs to be deposited in a malignant or inflammatory tissue, radionuclides emitting low-range highly-ionising radiation, i.e. α or β− particles, conversion and/or Auger electrons, are of great interest. Among the β− emitters, the radionuclides 32P (T ½ = 14.3 days), 89Sr (T ½ = 50.5 days), 90Y (T ½ = 2.7 days), 131I (T½ = 8.0 d), 153Sm (T½ = 1.9 d), 169Er (T½ = 9.4 d), 177Lu (T ½ = 6.7 days) and 188Re (T ½ = 17.0 h) are often used. The radionuclides 90Y and 188Re are available through generator systems whereas the others are produced directly or via decay of the respective precursors. Regarding α-emitting radionuclides, hitherto 211At (T ½ = 7.2 h) and 223Ra (T ½ = 11.4 days) have found some application but several other radionuclides are in development (see below). As far as Auger electron and X-ray emitters are concerned, so far only 125I (T ½ = 59.4 days) and 103Pd (T ½ = 17.0 days), respectively, have found wide application. The radionuclide 131Cs (T ½ = 9.7 days) is also finding increasing application.

The radionuclides 211At and 103Pd are produced at cyclotrons. All of the other radionuclides are produced via irradiations in a nuclear reactor. Except for some small discrepancies in a few cases, the decay and production data of all those radionuclides are fairly well known [cf. 3, 14].

Standardisation of nuclear data

When enough data have been accumulated for a certain radionuclide, it is recommended to evaluate and standardise the data to assist the user to choose proper conditions for the production of a medical radionuclide with acceptable quality.

As regards commonly used reactor-produced radionuclides, extensive neutron-induced reaction cross section data are available (e.g. in ENDF/B-VII.0), thanks to energy-related reactor programmes. Of special interest are (n,γ) reaction cross sections and fission yields. For some special cases, double neutron capture and (n,p) reaction cross sections are also needed and are available in the data files.

For accelerator-based radionuclides, the quality of available data varies. In case of photon-induced reactions, e.g. (γ,xn) and (γ,xp) reactions, an evaluated data file was prepared [8], but it is mostly based on nuclear model calculations. More work is needed to develop this file further.

As regards charged-particle induced reaction cross section data, the evaluation aspect was partly discussed earlier [4]. Considerable experimental efforts have been devoted in recent years by many groups. All data are promptly compiled in the EXFOR file. However, no Data Centre or research group was doing evaluation of charged-particle data. Therefore, the IAEA embarked on this mission about 20 years ago and organised two successive coordinated research projects (CRPs) in which about a dozen laboratories participated, under the guidance and chair of this author. Since no evaluation methodology existed at the start of the first project in 1995, the initial work was rather empirical. However, in later years strong application of nuclear models could be built in. All the major available nuclear model codes, namely GNASH, STAPRE, ALICE-IPPE, TALYS and EMPIRE were used.

While developing the evaluation methodology, three points clearly emerged:

  1. (a)

    the need to adjust the reported formation sections of a radionuclide according to its latest available decay data, especially γ-ray intensities,

  2. (b)

    the absolute necessity to standardise reactions used for charged-particle flux measurement. This need led to the development of a large number of monitor reactions for proton, deuteron, 3He- and α-particle beams over a wide range of energies [9]. In fact through the use of those monitor reactions, the accuracy of data reported in recent years from various laboratories has considerably increased.

  3. (c)

    the nuclear model calculations reproduce the experimental data with varying degrees of success. Their main objective in the evaluation process should be to check the consistency in the reported data.

The final recommended excitation function is generally based on statistical fitting of the selected experimental data. The evaluated and recommended data for the major diagnostic radionuclides are now available in [9] and those for the therapeutic radionuclides in [14]. They should allow a proper selection of the projectile energy range in a target to ensure high radionuclidic purity of the desired radionuclide, and the accompanying radioactive impurities can be accurately calculated. The two files are being continuously updated. The discrepancies in the data for the 18O(p,n)18F reaction found in the file [9] were removed through extensive new measurements [13]. Similarly new experiments were done to remove the discrepancy in the natRb(p,xn)82Sr reaction data [10] as well as in the natGa(p,xn)68Ge reaction data [11]. Recently the IAEA has initiated a third CRP to deal with other emerging radionuclides. A few independent research groups are also engaged in the evaluation work [cf. 15].

Alternative routes of production of 99Mo and 99mTc

The future supply of 99mTc (via the fission-produced 99Mo parent) appears to be somewhat jeopardized. On one hand the presently used reactors are ageing and there appears to be no planning for their replacement and, on the other, there is the risk of nuclear weapons´ proliferation due to the use of highly enriched 235U as the target material. Both these aspects demand a search for alternative methods of production and separation of 99Mo and/or 99mTc. Presently considerable efforts are underway in this direction.

The suggested routes for the production of 99Mo and 99mTc which are not dependent on a nuclear reactor are given in Table 1. Although the cross sections of those reactions [16–24] are not small, the non-availability of high fluxes of the respective projectiles leads to the conclusion that presently only the direct production of 99mTc via the 100Mo(p,2n)99mTc reaction is promising. The 100Mo(d,3n)99mTc reaction [20] is also interesting but, due to less availability of high-intensity deuteron beams, the emphasis is presently on the (p,2n) process.

Table 1 Alternative routes for production of 99Mo and 99mTc
Full size table

A study of the 100Mo(p,2n)99mTc reaction done at Jülich at the request of the IAEA had shown that this route using 22 MeV protons on a highly enriched 100Mo target is worth considering for local production and consumption of 99mTc [21]. Over the last few years a large number of other measurements have been reported but the data show large scatter, especially over the energy range of 10–20 MeV. The values from Prague [22], Edmonton [23] and Debrecen [24], however, appear to be more accurate. Recently a critical evaluation of all available data was carried out [cf. 25] using a rigorous nuclear model analysis and statistical fitting procedure; the recommended excitation function [25] is shown in Fig. 2. This should allow calculation of the yield of 99mTc with enhanced confidence. Over the suitable energy range of E p = 22 → 10 MeV, the calculated yield of 99mTc amounts to 700 MBq/µAh.

Fig. 2
figure 2

Excitation functions of the 100Mo(p,2n)99m,gTc and 100Mo(p,3n)98Tc reactions. The bold curve gives the evaluated values for the 100Mo(p,2n)99mTc reaction [25]. The experimental data for the 100Mo(p,2n)99gTc reaction were taken from [23] and the calculated data (using the codes TALYS and/or STAPRE) for the formation of 99gTc and 98Tc are from [25]

Full size image

It should be pointed out that the long-lived impurities 99gTc (T ½ = 2.1 × 105 a) and 98Tc (T ½ = 4.2 × 106 a) associated with the 100Mo(p,2n)99mTc process for the production of 99mTc may have strong repercussions on the utilization of 99mTc for medical applications. The theoretically calculated excitation functions for the formation of both those impurities are shown in Fig. 2 together with the recommended curve for the 100Mo(p,2n)99mTc reaction. A recent mass spectrometric measurement of 99gTc [23] confirms the calculation up to 18 MeV. The cross section for the formation of 99gTc is about four times higher than that for 99mTc. From about 17 MeV onwards, the formation of 98Tc also becomes significant. Evidently, even while using 100 % enriched 100Mo as target material, the choice of the incident proton energy will have to be carefully controlled to keep the level of the metal technetium impurity low so that 99mTc of the desired specific activity is obtained. This aspect may turn out to be critical in the acceptance of cyclotron-produced 99mTc for medical application. Thus, further extensive work is needed regarding the associated radionuclidic and metallic impurities.

Research oriented radionuclides

In recent years, nuclear data research related to medical radionuclides has been carried out mainly in two directions:

  • Development of novel positron emitters

  • Development of novel highly-ionising radiation emitters for internal radiotherapy

A brief discussion of those two areas is given below.

Novel positron emitters

With the growing significance of PET in diagnostic nuclear medicine, the need for novel positron emitters, also known as non-standard positron emitters, has been increasing, especially for studying slow metabolic processes and for quantification of targeted therapy.

The decay data of the emerging positron emitters are fairly well known; yet some discrepancies exist in a few cases, especially with regard to the β+ emission intensities. Since in recent years many of the novel positron emitters can be obtained in a radionuclidically pure form, a combination of beta counting and high-resolution X-ray spectrometry can lead to very reliable results. Using this approach, recently β+ emission intensities were accurately determined for 120I [26], 124I [27], 64Cu and 76Br [28]. In several other cases as well such measurements are recommended [cf. 7].

As regards production data of novel positron emitters, the subject has been treated in detail in two recent reviews [29, 30]. In this article, therefore, only a few salient features are discussed. The development of a suitable production process of a novel positron emitting radionuclide demands investigation of several nuclear reactions. In most cases it is necessary to use a highly enriched isotope as target material. The choice of the production route then depends mainly on the yield of the desired radionuclide and the expected level of the radioactive impurities. Other factors to be taken into account are the cost of the enriched target material, the ease of chemical separation of the product and the recovery of the enriched target isotope. The availability of the charged particles and their energies are also important. All those factors are well illustrated through measurements done in connection with the production of 124I (T ½ = 4.18 days), mostly at Jülich in collaboration with the groups at Brussels, Cairo, Debrecen and iThemba LABS (Cape Town). The results are summarized in Table 2. The yields were calculated from the single measurements [31–33] on the (p,n), (d,2n) and (p,2n) reactions, but from the evaluated data for the other reactions [34, 35]. Of all the processes investigated, the 124Te(p,n)124I reaction is the method of choice. Although the yield of 124I is not very high, the product obtained is of the highest radionuclidic purity. It has the added advantage of being used at a small-sized cyclotron. This reaction was originally suggested by Scholten et al. [31] and it is now universally applied for the production of 124I.

Table 2 Routes for production of 124I. The 124Te(p,n)124I reaction gives the purest form of 124I
Full size table

Most of the novel positron emitters have been developed using a small-sized cyclotron (with E < 20 MeV). A brief overview of the production routes of several emerging positron emitters is given in Table 3. The common route of production of each radionuclide is the low-energy (p,n)-reaction on an enriched target isotope. In a few cases other low-energy reactions, such as (d,n) and (p,α), have also been employed. For a detailed list of references relevant to nuclear data measurements see a recent review article [30]. For all the listed radionuclides some sort of novel medical application has been reported. The radionuclides 64Cu, 86Y, 89Zr and 124I, on the other hand, have already found wide application. The number of relevant reactions studied for their production has therefore also been correspondingly large. However, the low-energy (p,n)-reaction on an enriched target isotope is favoured because it generally leads to a high-purity product.

Table 3 Some novel positron emitters for medical applications
Full size table

Figure 3 depicts the excitation function of the 64Ni(p,n)64Cu reaction which may be regarded as a typical (p,n) reaction in the medium mass region. It was used first at Jülich [36] for the production of 64Cu and is now the method of choice at almost all laboratories. A later publication [37] supported those data. The excitation function can be described relatively well by nuclear model calculations [38].

Fig. 3
figure 3

Experimental data along with the results of nuclear model calculations using the codes EMPIRE, STAPRE and TALYS for the 64Ni(p,n)64Cu reaction. The encircled data were deselected in further evaluation (taken from Ref. [38])

Full size image

It should be pointed out that three of the novel positron emitters listed in Table 3, namely 86Y, 94Tc and 120I, have isomers. The choice of the nuclear reaction is then critical to obtain one isomer with the least possible amount of the other isomer (for more discussion [cf. 2]). Incidentally, in all the three cases, the desired radionuclide is a low-spin isomer. So the low-energy (p,n) reaction on the corresponding highly enriched target, viz. 86Sr(p,n)86Y [39], 94Mo(p,n)94mTc [40] and 120Te(p,n)120gI [41], gave the purest form of the desired product.

In addition to the positron emitters listed in Table 3, nuclear data work on several other potentially useful positron emitting radionuclides, e.g. 44Sc (T ½ = 3.9 h), 45Ti (T ½ = 3.1 h), etc. is also going on, especially in the energy region below 20 MeV. However, many useful or potentially useful positron emitters can be produced only using intermediate energy reactions. Extensive nuclear data work has shown that the production of the radionuclides 52Fe (T ½ = 8.3 h), 73Se (T ½ = 7.1 h), 77Kr (T ½ = 1.2 h) and 83Sr (T ½ = 32.4 h) demands a high intensity cyclotron or accelerator, delivering protons of energies up to about 70 MeV (in the case of 52Fe preferably up to 100 MeV).

Besides protons, heavier charged particles like deuterons, 3He- and α-particles may also induce a few useful reactions. For example, intermediate energy deuterons could be useful in the production of 64Cu via the natZn(d,x)-process, a 3He-particle beam in the production of 75Br by the 75As(3He,3n)-process and an α-particle beam in the production of 73Se via the 70Ge(α,n)73Se reaction (for more details cf. [30]). Several newer studies are in progress in a few laboratories (for example, see the contribution by Ditroi et al., this issue of the journal).

Similar to some commonly used positron emitters made available via generator systems, a few novel and potentially useful positron emitters could also be produced through generators. Some of them are: 44Ti (60.4 a)/44Sc (3.9 h), 72Se (8.5 days)/72As (26.0 h), 140Nd (3.4 days)/140Pr (3.4 min), etc. All of the above mentioned parent radionuclides can be produced only using intermediate energy charged particle beams.

Novel therapeutic radionuclides

The number of potentially interesting therapeutic radionuclides is very large. However, as mentioned above, in internal radionuclide therapy the emphasis has shifted to low-range but highly-ionising radiation emitters. In a recent review [4] many of the novel therapeutic radionuclides were discussed in some detail. The scope of this article is limited to the radionuclides which are presently attracting most attention. Their decay characteristics are summarized in Table 4. The decay and production data of each radionuclide are considered below individually.

Table 4 Decay characteristics of some important emerging therapeutic radionuclides
Full size table

67Cu

This radionuclide is gaining increasing significance as a theragnostic pair of the β+ emitter 64Cu. Its decay data are well known and the production methods have been recently reviewed [42]. Some of the reaction data have also been evaluated [14]. In general six reactions could possibly be used: (a) 70Zn(p,α)67Cu, (b) 70Zn(d,nα)67Cu, (c) 68Zn(p,2p)67Cu, (d) 64Ni(α,p)67Cu, (e) 68Zn(γ,p)67Cu, (f) 67Zn(n,p)67Cu. The first reaction has found practical application [cf. 43] but the yield was rather low. The 68Zn(p,2p)67Cu reaction appears to be the most promising. The measured reaction cross sections, however, show some discrepancies [44–46]. Nonetheless, it is a relatively high-yield process and 67Cu could be produced in batches of a few GBq [cf. 47] using a high-current enriched 68Zn target and a proton energy range of E p = 80 → 30 MeV. As regards the 70Zn(d,nα)67Cu and 64Ni(α,p)67Cu reactions, so far only cross sections have been measured [48, 49] and the expected respective yield is low. In recent years there has been an increasing interest in the utilization of electron linear accelerators for medical radionuclide production and the reaction 68Zn(γ,p)67Cu has been investigated in some detail [50]. The yield of 67Cu for a natZn target amounting to 1 MBq/(g kW h) is, however, rather low. Furthermore, even the old method used at nuclear reactors, namely the 67Zn(n,p)67Cu reaction, has received some newer attention [51]. While measuring cross sections of the 64Zn(n,p)64Cu and 67Zn(n,p)67Cu reactions in a TRIGA reactor spectrum, Uddin et al. [51] came to the conclusion that for theragnostic studies a good mixture of 64Cu and 67Cu could be produced. The cross sections of the two reactions differ considerably, possibly due to a significant difference in the proton binding energies in the two target nuclei [52]. The calculated yield of 67Cu amounts to 4.4 MBq/(g h for 1014 n cm−2 s−1). On the other hand, it remains questionable whether one can achieve in reactor production the same chemical and radiochemical purity of 67Cu as in cyclotron production. The emphasis should therefore be on the 68Zn(p,2p)67Cu process using intermediate energy protons.

186Re

Since the chemistry of Re is similar to that of Tc and since a large number of Tc-compounds are already finding applications in nuclear medicine, the radionuclide 186Re is attractive for internal radiotherapy. The γ-ray emission intensities in its decay have been re-measured [53]. The original method of its production, namely the 185Re(n,γ)186Re reaction, has been abandoned due to the resulting low specific activity. The new method of preparation of this radionuclide is the 186W(p,n)186Re reaction, though the 186W(d,2n)186Re reaction has also been investigated and gives a higher yield than the (p,n) reaction. Most of the studies deal with nuclear reaction cross sections [cf. 54–59]; only in a few papers small scale production has been reported. A recent critical analysis [60] showed that, for obtaining a high-purity product, an enriched 186W target is absolutely necessary and the maximum proton energy should not exceed 18 MeV. The formation of the very long-lived isomer, 186mRe (T ½ = 2 × 105 a), also needs special consideration in both (p,n) and (d,2n) reactions. Based on nuclear model calculations the specific activity of 186Re was also predicted [60]. More recent experimental work in this direction [61, 62] appears to support the theoretical considerations. Presently, some practical yield measurements [cf. 63] and other efforts are underway in a few laboratories to produce this radionuclide in larger quantities.

225Ac

There is great interest in this α-particle emitting radionuclide; it is useful in itself as well as for providing 213Bi (T ½ = 46 min; E α = 5,900 keV) through a generator system. The decay data involved in the relevant decay chain have been evaluated [cf. 14] and the half-life of 225Ac was recently re-determined [64]. Regarding its production, presently extensive effort is being invested. On one hand its separation from nuclear waste (229Th) is being optimised [65] and, on the other, the 226Ra(p,2n)225Ac reaction, making use of the radioactive target material 226Ra, is being developed [cf. 66]. A third possibility under investigation is its production via irradiation of 232Th with intermediate energy protons [67–69].

The cross sections for the formation of 225Ac in the proton irradiation of 232Th [67, 68] are shown in Fig. 4. The calculated 225Ac yield over the energy range E p = 140 → 60 MeV amounts to 4 MBq/µA h. The amount of 227Ac impurity in 225Ac was also estimated by Ermolaev et al. [67] and the result is given in Fig. 4. Over the above mentioned energy range the level of the 227Ac impurity amounts to about 0.2 % of 225Ac. It should be mentioned that the yield of 225Ac via the 226Ra(p,2n)-reaction over the energy range E p = 22 → 10 MeV amounts to 7 MBq/µA h. The yield through the 232Th(p,x)225Ac reaction is thus lower and the chemical processing involved is extensive. But it avoids the use of a radioactive target which is quite cumbersome. Anyway, all methods of 225Ac production need further development. In fact extensive efforts are underway with regard to the production of 225Ac via both the latter processes.

Fig. 4
figure 4

Cross section data for the formation of 225Ac in proton-induced reactions on 232Th (left scale) shown as a function of projectile energy [taken from Refs. 67, 68]. The thick solid line is an eye-guide. The level of the 227Ac impurity (thin solid line, right scale) is also given [67]

Full size image

117mSn

This low-lying isomeric state of 117Sn decays by heavily converted isomeric transition to the ground state. The emitted conversion electrons of energies about 160 keV are ideally suited for some internal therapeutic studies and the radionuclide has been under consideration for more than three decades. It has been produced in small quantities via four routes: (1) 117Sn(n,n´γ)117mSn, (2) 121Sb(p,2p3n)117mSn, (3) natCd(α,xn)117mSn, (4) natIn(α,pxn)117mSn. Very recently in a joint effort between BNL and Clear Vascular, Inc. it has been realized that the product obtained through the first two processes does not have a specific activity (and also chemical purity) high enough for medical applications. The more promising method for production appears to be the α-particle irradiation of cadmium or indium.

The cross section data for the 115In(α,d + pn)117mSn reaction were determined by Fukushima et al. [70] up to 40 MeV and by Qaim and Döhler [71] up to 140 MeV. In both the works, a peak at about 38 MeV was observed; the magnitude of the cross section in the latter work was, however, by about a factor of 5 higher. From the data of Qaim and Döhler [71], the optimum energy range for the production of 117mSn was deduced to be E α = 45 → 20 MeV, with the thick target yield of 117mSn as 370 kBq/µAh and the 113Sn (T ½ = 115.9 days) impurity level of <0.2 %. The yield is rather low but the advantage is that the main target isotope 115In has an abundance of 95.7 % in natIn so that an enriched target is not necessary.

The cross section data for the 116Cd(α,3n)117mSn reaction were measured by Montgomery and Porile [72] up to 45 MeV and for the natCd(α,xn)117mSn process by Qaim and Döhler [71] extensively over the energy range of 15–140 MeV. The excitation curve is composed of two reactions, namely 114Cd(α,n)117mSn and 116Cd(α,3n)117mSn. The latter reaction has higher cross section and is more important. The relatively high formation cross section of 117mSn in α-particle induced reactions is attributed to its high spin (11/2−), which is expected to be preferentially populated [cf. 73–75].

In two more recent measurements under Brussels/Debrecen cooperation the reactions natCd(α,x)117mSn, 114Cd(α,n)117mSn and 116Cd(α,3n)117mSn were investigated up to 39 MeV [76, 77]. The data for the natCd(α,x)117mSn reaction [76] are slightly higher than those by Qaim and Döhler [71]. However, if the data of Qaim and Döhler [71] for the peak around 45 MeV, which is of prime interest in the production of 117mSn, are normalised to 100 % 116Cd enrichment (instead of 7.5 % in natCd), the resulting cross section values of about 1,200 mb are comparable to those reported by Adam Rebeles et al. [77] as well as by Montgomery and Porile [72]. Based on the data by Qaim and Döhler [71], which cover the full energy range of the peak, the energy region for the production of 117mSn via the 116Cd(α,3n)-reaction is therefore suggested as E α = 60 → 30 MeV. The thick yield of 117mSn for a 100 % 116Cd target would amount to about 8.4 MBq/µAh with almost no isotopic radionuclidic impurity. Due to this reason the Clear Vascular, Inc. has chosen this process for production purposes. In view of the importance of this reaction it is recommended that more cross section measurements be done up to 70 MeV using highly enriched 116Cd as target material.

193mPt and 195mPt

These two radionuclides are pure X-ray and Auger electron emitters, each decay leading to more than 30 secondary electrons, with their energies distributed between 10 and 130 keV. Since platinum-complexes (like cis-di-chlorodiaminplatinum) have been in use in chemotherapy as potent anti-tumour agents for a long time, both 193mPt and 195mPt have great potential in Auger electron therapy. So far the major drawback in their widespread use has been their non-availability with a high specific activity. The nuclear reactions employed, namely 192Pt(n,γ)193mPt and 194Pt(n,γ)195mPt, even while using highly enriched isotopes as target materials, do not lead to specific activity high enough for therapeutic application. Attempts to produce 195Ir via double neutron capture by 193Ir and its subsequent decay to 195mPt (i.e. via the route \( ^{{193}} {\text{Ir}}\left( {{\text{n}},\gamma } \right)^{{194}} {\text{Ir}}\left( {{\text{n}},\gamma } \right)^{{{\text{195m,g}}}} {\text{Ir}}{\mathop{\xrightarrow{{\beta ^{ - } }}}}{^{{{\text{195m}}}} {\text{Pt}}} \)) have not been very successful. Due to the very high spin (13/2+) of both 193mPt and 195mPt, it was considered more fruitful to investigate the α-particle induced reactions on highly enriched 192Os. In two recent studies involving clean radiochemical separation of radioplatinum followed by X-ray spectrometry the excitation functions of the reactions 192Os(α,n)195mPt [78, 79] and 192Os(α,3n)193mPt [78, 79] could be measured. The results are shown in Fig. 5. In small scale production runs it could also be demonstrated that both the radionuclides can be produced with high specific activity; the achieved yield of 195mPt is rather low but that of 193mPt over the energy range E α = 40 → 30 MeV amounts to about 10 MBq/µA h. Thus this radionuclide can be produced in quantities sufficient for therapeutic applications.

Fig. 5
figure 5

Excitation functions of the 192Os(α,n)195mPt and 192Os(α,3n)193mPt reactions determined radiochemically using X-ray spectrometry [78, 79]. The solid lines are based on nuclear model calculations [cf. 79], and the dashed line is an eye-guide

Full size image

Future perspectives and concluding remarks

The foregoing discussion shows that nuclear data play a very important role in the development of radionuclides for applications. The radioactive decay data govern the choice of a radionuclide for medical application and the nuclear reaction data allow optimisation of its production route. Whereas the decay data are well known, except for some small discrepancies in a few cases, the production data need more attention. In the case of neutron-induced reactions, cross sections are needed for radionuclide production in a nuclear reactor, and extensive information is available. However, regarding photon- and charged-particle induced reactions, further research and development work is necessary. The new data needs generally refer to development of novel positron emitters and low-range highly-ionising radiation emitters for internal radiotherapy. Whereas the former are generally produced at a low-energy (E < 20 MeV) cyclotron via the low-energy (p,n), (p,α) and (d,n) reactions, etc., for the production of the latter a medium-sized cyclotron is required. On the other hand, for production of several positron emitters, like 52Fe (T ½ = 8.3 h), 73Se (T ½ = 7.1 h), 83Sr (T ½ = 32.2 h), etc. and therapeutic radionuclides such as 67Cu (T ½ = 2.58 days), 225Ac (T ½ = 10.0 days), etc., intermediate energy protons from a cyclotron or an accelerator are needed. Another point worth considering is the availability of multiparticle cyclotrons. The deuteron-induced production of a few radionuclides leads to higher yield than the proton-induced production, e.g. 64Cu (T ½ = 12.7 h) from a 64Ni target, 103Pd (T ½ = 17.0 days) from a 103Rh target, 186Re (T ½ = 3.7 days) from a 186W target, etc. The available deuteron beam currents are, however, generally much lower than for proton beams. On the other hand, a few modern cyclotrons/accelerators can deliver high-intensity deuterons (e.g. ACSI TR30 in Canada, SARAF in Israel) so that the use of deuteron-induced reactions may become advantageous. The α-particle beam is also very useful in populating low-lying high-spin isomeric states [cf. 73–75] which deexcite by IT and emit low-energy electrons (conversion or Auger, or both) that can be used in internal therapy. Examples are 117mSn, 193mPt, 195mPt, etc. The cross section data measurements in those cases are more challenging because often X-ray spectrometry is required. The increasing future demands of the novel radionuclides will decide what type of cyclotrons or accelerators should be favoured and constructed. The emphasis in nuclear data work will also have to be adjusted accordingly.

Over the last few years, two important new developments in the application of radionuclides have been emerging. The first one deals with a combination of PET and internal radiotherapy (as first demonstrated by Herzog et al. [80]), and the second one with a combination of PET and magnetic resonance imaging (MRI). In the latter case a β+ emitting radionuclide of the contrast agent manganese [e.g. 51Mn (T ½ = 46.2 min) or 52Mn (T ½ = 5.6 days)] would be useful. The production data for those radionuclides have been recently measured [cf. 81, 82]. Evidently, for those two fast expanding applications, further novel radionuclides will have to be developed (for more details cf. [4]). On a longer term basis, the radionuclide research could also involve a combination of radioactivity and nanotechnology, provided toxicity problems are overcome. The future perspectives of radionuclide research for medical applications thus appear to be very bright. In fact it is a flourishing and fast expanding field, even in developing countries. The supporting nuclear data research will therefore also continue to receive further attention.