Production, processing and small animal PET imaging of titanium-45

Nuclear Medicine and Biology 32 (2005) 117 – 122 www.elsevier.com/locate/nucmedbio

Production, processing and small animal PET imaging of titanium-45 Amy L. Va¯verea,b, Richard Laforesta,c, Michael J. Welcha,b,c,* a

Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO 63110, USA b Department of Chemistry, Washington University in St. Louis, St. Louis, MO 63104, USA c Alvin J. Siteman Cancer Center, Washington University School of Medicine, St. Louis, MO 63110, USA Received 9 August 2004; received in revised form 1 October 2004; accepted 18 October 2004

Abstract Introduction: Titanium-45 was prepared as a tool for elucidation of the mechanism of action of titanium anticancer drugs in vivo using microPET imaging. Methods: Titanium-45 was produced by the 45Sc(p,n)45Ti nuclear reaction using 14.5 MeV protons. Sufficient yields of 45 Ti were produced and separated from the target material with 99.8% radionuclidic purity using a simple, efficient separation procedure. Results: A typical bombardment of 5 AA for 1 h produced an average of 2105F150 MBq (56.9F4.0 mCi) at the end of bombardment (EOB), well within acceptable range of the calculated theoretical yields of 2165 MBq and 433 MBq AA–1 h–1 (58.5 mCi and 11.7 mCi AA–1 h–1). This amount of activity is sufficient for the radiosynthesis of target compounds as well as imaging studies. MicroPET images of a miniature Derenzo phantom show excellent resolution where rods of 1.25 mm were separated by four times their diameter. Conclusions: Titanium-45 can be easily produced on a biomedical cyclotron with excellent yields as compared to calculated theoretical values with imaging studies demonstrating that the decay properties of titanium-45 are well suited for microPET. D 2005 Elsevier Inc. All rights reserved. Keywords: Titanium-45; MicroPET; Ion chromatography

1. Introduction Over the past several years, there has been substantial interest in the application of titanium complexes as anticancer drugs. Titanium(IV) complexes, particularly titanocene complexes, have been shown to exhibit high antitumor activity against a range of tumors in animals with less toxic side effects than cisplatin [1– 4]. One of these compounds, titanocene dichloride, is currently in phase II clinical trials as an anticancer agent [5,6]. Although the aqueous chemistry of titanium compounds has been explored [7], little is known about the biological chemistry of these agents. Under physiological conditions, titanium(IV) is readily taken up by human transferrin from titanocene dichloride [8], suggesting that this protein may mediate the uptake of titanium from anticancer drugs.

* Corresponding author. Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO 63110, USA. Tel.: +1 314 362 8435; fax: +1 314 362 9940. E-mail address: [email protected] (M.J. Welch). 0969-8051/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.nucmedbio.2004.10.003

Titanium-45 has a half-life of 3.09 h and decays 85% by positron emission and 15% by electron capture with an E h+max = 1.04 MeV [9]. Positron decay of titanium-45 proceeds mostly to the ground state of scandium-45, so almost no other concurrent gamma rays are emitted (less than 0.1%) apart from the two-annihilation photons. The maximum energy of the emitted positron is 1040 keV with an average value of 439 keV, values that are similar to carbon-11 (E max =960 keV, E avg =386 keV). This relatively low maximum positron energy and high positron emission percentage make 45Ti an ideal candidate for positron emission tomography (PET). Due to the increased resolution afforded by small animal PET cameras, the properties of the individual isotopes have a more pronounced affect on image quality. In a recent study [10], the maximum positron range was measured for several clinically used and nonconventional nuclides. Oxygen-15, routinely used as H215O on clinical PET scanners, had a maximum positron range of 5.8 mm and an average range of 2.28 mm. This range has little effect when imaging using human PET scanners, but the image degradation caused by this range is too great for microPET imaging. Titanium-45

118

A.L. Va¯vere et al. / Nuclear Medicine and Biology 32 (2005) 117–122

was calculated to be 3.04 mm, with an average range of 1.14 mm. This is comparable to the widely used radionuclide 11 C whose maximum and average positron ranges reported in the same study were 2.7 and 0.98 mm, respectively. The preparation of 45Ti on low energy cyclotrons has been described using proton bombardment of natural scandium [11,12]. Bombardment times and currents, and resulting yields were not reported. In this procedure, a natural scandium foil was irradiated with protons in a cyclotron inducing a 45Sc(p,n)45Ti nuclear reaction. Once irradiated, the target was dissolved in 6.0 N HCl to give [45Ti]Cl3 to which a drop of concentrated nitric acid is added to oxidize the Ti(III) to the +4 state. The HCl was then removed by evaporation and the residue is redissolved in HCl three times to remove all nitric acid. The 45Ti is then redissolved in 6 N HCl and separated from the scandium by elution down a cation-exchange column. The scandium was recovered by a final elution of the column with 4 N HCl containing 0.1 N hydrofluoric acid. The repeated evaporation steps in this procedure are time-consuming; therefore, the development of a simplified procedure was pursued. Published work on the applications of 45Ti has been limited, but this nuclide has been explored as a potential metal for labeling pharmaceuticals for PET imaging. In 1982, the use of 45Ti as a possible imaging agent was proposed by labeling human serum albumin, phytate, DTPA and citrate [12]. When injected into rats, [45Ti]DTPA crossed the blood–brain barrier. Uptake in the spleen, lung and liver was observed with the [45Ti]phytate. Following this initial report, an autoradiographic study explored [45Ti]ascorbic acid (AsA) as a possible radiopharmaceutical by studying its biochemical behavior in rats [13]. Results demonstrated that [45Ti]AsA likely formed a complex with albumin, unlike the coinjected [14C]AsA. The same researchers later reported biodistribution studies of the same complexes [14]. High liver and spleen uptake were observed, and AH109A tumor uptake of both DTPA and citrate complexes in rats was lower than blood uptake. These results also suggested that ligands of these complexes were displaced by plasma proteins or that the complexes themselves bound to proteins. In this report, we present data necessary to pursue 45Ti as a tool for the examination of titanium biodistribution using MicroPET. Previously reported production and processing techniques for 45Ti were simplified. A targetry system initially developed for the production of 64Cu [15] was adapted to the production of a variety of other metal radionuclides, including 45Ti. This targetry system was modified for the use of a foil target holder previously reported for the production of 66Ga [16], and was shown to be adequate for efficient production of 45Ti. The use of 45Ti for imaging with MicroPET was also explored. 2. Materials and methods All chemicals, unless otherwise stated, were purchased from Aldrich (Milwaukee, WI). All solutions were prepared

using distilled, deionized water (Milli-Q; N18 MV resistivity). Radioactive samples were counted in a Capintec dose calibrator (Capintec CRC-10R, Capintec, Ramsey, NJ) and in a Beckman 8000 g-counter (Beckman Instruments, Irvine, CA). Radionuclidic purity was determined by analysis with a Canberra multichannel analyzing germanium gamma spectrometer (Meriden, CT). Infrared spectra were obtained on a Perkin–Elmer Spectrum BX FT-IR Spectrometer (Wellesley, MA) using diffuse reflectance sampling accessories for the solid samples and injection between NaCl plates for solution samples. Mass spectrometry was performed using a Waters ZQ 4000 mass spectrometer (ESI +ve mode; 130–800 Da range). Small animal PET imaging was performed on a Concorde Microsystems MicroPET-R4 (Knoxville, TN) [17]. 2.1. Titanium-45 production and processing Titanium-45 was produced and processed using methods similar to that of Ishiwata et al. [14], employing a 45 Sc(p,n)45Ti nuclear reaction. In a glove box under nitrogen atmosphere, a 77-mm2 natural scandium foil (0.250 mm thick, 100% abundant, Alfa Aesar, Ward Hill, MA) was cut and screwed into an aluminum target holder previously described for the production of 66Ga [16], designed in house (R.L.) and fabricated by the machine shop at Washington University School of Medicine. The target was irradiated with 14.5 MeV protons at 3–10 AA for 30 min to 2 h using the Cyclotron (Berkeley, CA) CS-15 biomedical cyclotron at Washington University. High-pressure water flowing through a depression on the back of the target holder provided efficient cooling during irradiation. After bombardment, the irradiated scandium foil was dissolved in 2 ml of 6.0 N HCl (99.999999% pure, Alfa Aesar) and applied to a cation-exchange column containing AG 50W-X8 resin (1.913 cm), 100 –200 mesh (Bio-Rad Laboratories, Hercules, CA), conditioned with 6.0 N HCl. Seven 6-ml fractions were collected and counted to determine percent recovery of radioactivity. The radionuclidic purity was determined by analysis with a multichannel gamma spectrometer. The [45Ti]/HCl solution was evaporated to dryness by heating under a stream of nitrogen. The 45Ti residue was then ready to be dissolved in the desired solvent for further studies. The theoretical production yields were calculated using published (p,n) reaction cross sections integrated over the energy of the impinging protons as they are slowed down in the target foil. Published cross-sectional values were retrieved from EXFOR utility maintained at Brookhaven National Laboratory. The total excitation function between 3.28 and 19.3 MeV was obtained by combining the experimentally measured production yields from the publications of Dell et al. [18] (3.28 –6.77 MeV) and Levkovskij [19] (7.7–19.3 MeV). The energy loss of the protons while passing through the target material is calculated from TRIM stopping power tables. Calculations

A.L. Va¯vere et al. / Nuclear Medicine and Biology 32 (2005) 117–122

119

for the (p,2n) reaction were also performed using crosssectional data from Levkovskij.

3. Results

2.2. Ion chromatography

To date, over 75 productions totaling 115 Bq (3.1 Ci) of Ti have been carried out by the method reported herein with an average overall yield of 422F30 MBq AA–1 h –1 (11.4F0.8 mCi AA–1 h–1). A typical bombardment of 5 AA for 1 h produced an average of 2105F150 MBq (56.9F4.0 mCi) at end of bombardment (EOB) (Table 1). Calculated theoretical yields were 2160 MBq and 433 mCi AA–1 h–1 (58.5 mCi and 11.7 mCi/AAh). Melting of the aluminum target holder was caused at an irradiation of 10 AA and occasionally at 5 AA. After bombardment, complete processing of 45Ti and preparation of one fraction was achieved in 60 min or less resulting in an average of 370 MBq (10 mCi) of 45Ti for further use. Preparation of additional fractions added approximately 30 min per fraction to account for removal of the HCl by heat under a stream of nitrogen. The solution must be heated enough to accelerate evaporation, but be maintained below boiling so as to avoid bumping or volatilization of the radioactive sample. Upon processing, an average of 92.3% of the activity was recovered from the cation-exchange separation. It is possible to recover the scandium starting material by elution of the column with 4 M HCl containing 0.1 M HF [14]; however, this was not performed due to the extremely low cost of target material.

Once the 45Ti was isolated from the remaining scandium starting material by cation-exchange chromatography, the collected fractions were analyzed by ion chromatography using a Waters ActION Analyzer (Milford, MA). This system consisted of a Waters series 600 pump, a Dionex IonPac CS-5A (2504 mm) column, a Timberline PostColumn Reagent Delivery Module and a Waters 486 Tunable Absorbance Detector. The mobile phase was 0.2 N HNO3 with 0.2 mM PAR [4-(2-Pyridylazo)resorcinol, monosodium salt hydrate], 1.0 M hydroxylamine hydrochloride and 0.9 M NaOH comprising the postcolumn solution. The flow rate was set at 0.5 ml min 1 with UV monitoring at 520 nm. Collected fractions of processed 45Ti were analyzed for scandium breakthrough, as well as for the presence of other metal contaminants. Samples were also analyzed days after all 45Ti had decayed and were compared to standards of scandium, titanium and iron to check for impurities and scandium breakthrough. A baseline chromatograph was taken using only 0.05 N HCl solvent, since 6 N would be too concentrated for the IC system. 2.3. Image quality A miniature Derenzo phantom [20] was used to evaluate Ti image quality by PET using the microPET-R4. The phantom was constructed from a Lucite cylinder (6 cm in diameter and 4 cm long) with holes drilled into it of diameters 1, 1.25, 1.5, 2 and 2.5 mm separated by four times their diameters. These rods were filled with an aqueous solution of approximately 500 ACi of activity, and data were collected from 20 to 45 min at 600 frames per second. Images were then reconstructed by Fourier Rebinning (FORE) followed by 2D-filtered back projection (2D-FBP, ramp filter at Nyquist frequency) using all tilt angles. Image quality was studied by visual inspection of transaxial slices in the middle of the phantom as well as profile analysis of the 1-mm rod images for 45Ti as compared to 18F. 45

2.4. Determination of chemical form Infrared spectroscopy was employed to aid the determination of the chemical form of the radioactive species of the processed titanium-45. To duplicate the species in a nonradioactive state, a 77-mm titanium foil (Alfa Aesar) was placed in 2 ml of 6.0 N HCl in an identical procedure used for processing the scandium target. The metal took a number of weeks to fully dissolve, but a pale purple solution was observed within 24 h. The solution was analyzed by mass spectroscopy and by IR spectroscopy by injection between NaCl plates, using 6 N HCl as the reference. The HCl was then evaporated under a stream of nitrogen with heating and the resulting residue was analyzed by IR using diffuse reflectance.

3.1. Titanium-45 production and processing 45

3.2. Radionuclidic purity The major impurities produced during the 3-AA, 45-min irradiations were measured by a gamma-ray spectrometer. In all cases, the radionuclidic yield of 45Ti was 99.8% with the only isotopic impurity being 55Co (t 1/2 = 17.5 h). Analysis of the aluminum target holder, which remained radioactive weeks after bombardment, revealed small amounts of several cobalt nuclides (55,56,57,58Co) most likely due to the target apparatus. Zinc-65 and 48V were also present on the target holder in small quantities, likely due to crosscontamination from the production of copper radionuclides using the same system. No measurable amount of Ti-44 was produced due to the unfavorable production threshold of 12.9 MeV and its long half-life of 47.3 h.

Table 1 Ti-45 production results Current (AA) 2 3 3 4 5 5 5 5 10

Time (min) 20 45 60 45 45 60 90 120 60

Ave. 45Ti at EOB (mCi)

Yield (mCi AA 1d h 1)

n

8.92 26.0F1.5 37.1 35.2 41.2F3.2 56.9F4.0 83.4 109.1 105.4

13.4 11.1F0.6 12.4 11.7 11.0F0.9 11.4F0.8 11.1 10.9 10.5

1 28 1 1 7 21 1 1 1

120

A.L. Va¯vere et al. / Nuclear Medicine and Biology 32 (2005) 117–122

Fig. 1. Ion chromatographs of

45

Ti fraction after decay (A) and 0.05 N HCl (B).

3.3. Ion chromatography Ion chromatographic analysis of the processed 45Ti fractions showed a radioactive peak at a similar retention time to that of the titanium standard at 10.34 min, with greater than 95% of the activity collected after elution. The processed fractions of 45Ti were also analyzed to assess scandium breakthrough or metal contamination. Analysis of the fractions exhibited the presence of two impurities (Fig. 1A) comparable to those seen in the chromatogram of the 0.05 N HCl sample (Fig. 1B).

revealing a width of 2.55 and 4.23 mm for 18F and 45Ti, respectively. The average peak to valley ratio can also be observed to decrease from 2.73 to 1.84 for these rods. 3.5. Chemical form The chemical form of the 45Ti species after processing has been postulated to be 45TiOCl2 [14]. In an effort to provide definitive proof, IR spectroscopy was performed on titanium samples prepared in the identical manner to the target processing method. Upon addition of 6 N HCl to the

3.4. Image quality The image quality of 45Ti was assessed by visual inspection of microPET images of the miniature Derenzo phantom. Clear resolution was observed down to a rod diameter of 1.25 mm (Fig. 2). This resolution is comparable to that of 18F, the most commonly used PET isotope, which can be resolved to a diameter of 1 mm. A slight degradation of spatial resolution can be seen due to the higher range of the 45Ti positrons. In Fig. 3, a profile was drawn through the 1-mm rods on the image. The first peak was fitted with a gaussian curve for each nuclide,

Fig. 2. Miniature Derenzo phantom imaged on a microPET-R4 using (A) 18 F and (B) 45Ti.

A.L. Va¯vere et al. / Nuclear Medicine and Biology 32 (2005) 117–122

Fig. 3. Profile distribution of Derenzo phantom image.

45

Ti and

18

F from 1-mm rods on a miniature

titanium foil, no change appeared to take place, but a pale purple color began to appear within 24 h, IR (NaCl) 3398, 1645 and 478 cm 1. By heating under a nitrogen stream, the HCl was removed from the titanium solution, resulting in a purple and white solid, IR (KBr) 3527, 2815, 2430, 1648 and 926 cm 1. MS results were inconclusive with no identification of titanium species present in solution. 4. Discussion Routine production of 45Ti by proton bombardment of a natural scandium foil resulted in good yields and high radiochemical purity. Use of the aluminum target holder provided efficient cooling for bombardments up to 5 AA, beyond which the target welded to the target holder. The targetry system routinely used at our institution for the production of copper radionuclides proved quite amenable to the production of other metal nuclides by bombardment of a solid target. The foil is maintained in close contact with the target holder only at the edge of the bombarded area. Good thermal contact with the cooled target holder is only achieved at the edge of the foil. The thermal conductivity of scandium is low (15.8 W m 1 K 1 at 25 8C) compared to aluminum (234.2 W m 1 K 1 at 25 8C) [21]. It is possible that such a weak thermal conductivity prevents efficient heat dissipation to the aluminum holder at the beam spot, as a result of energy degradation as the beam passes through the target material. Prolonged proton bombardments with high current beams may therefore result in melting of the target foil. Other techniques for foil bombardment are presently being investigated in order to achieve higher production rates. Fortunately, production yields of titanium-45 are high enough so that sufficient amount of activity can be produced in a reasonable time using this low beam current. Previously reported procedures for the separation of titanium-45 from the scandium starting material indicate that the addition of nitric acid ensures the +4 state of the titanium [14]. This does not appear to be necessary, as the metal is already oxidized upon introduction to the air and water in the

121

hydrochloric acid. Removal of the nitric acid addition step also eliminated the need for redissolving and evaporation of the target material in triplicate, which would add 45 min. Irradiation of a target material comprised of a single isotope, 45Sc, at 100% abundance resulted in high radionuclidic purity. The presence of the only impurity, 0.2% 55 Co, is most likely due to irradiation of trace nickel impurities in the aluminum target holder or from the targetry apparatus. It seems evident that no significant impurities were introduced either by the target material or the separation process. The recovery of the target material is not necessary due to the relatively low cost of the scandium foil. MicroPET images of the phantom filled with 45Ti were only slightly inferior to the 18F image due to its higher positron energy of 1.04 MeV as compared to 0.65 MeV for 18 F as observed visually and by profile analysis. The clear resolution of the 1.25-mm rods would easily lend this nuclide to microPET imaging of radiopharmaceuticals as well as for use on human PET scanners. Research groups have speculated on the chemical form of 45Ti after processing. Initially, it was identified as 45 TiCl4, but then reported to be most likely 45TiOCl2. Based on the simple chemistry of titanium in aqueous solution and hydrochloric acid, it is obvious that association with oxygen must occur. The chemical form also changes when the solution is evaporated to dryness, as shown by the IR data presented within. The stretch at 926 cm 1 in the solid compound is characteristic of a Ti= O stretch, typically found between 800 and 1100 cm 1 [22]. The broad stretch above 3000 cm 1 was most likely due to the presence of water in HCl in solution and possible formation of hydroxides in the solid state. Although oxygen association was evident, a definite identification of the processed species was not possible; therefore, further studies are necessary for the determination of the 45Ti species present. 5. Conclusion Titanium-45 can be easily produced on a biomedical cyclotron with excellent yields as compared to calculated theoretical values. Sufficient amounts were produced for labeling studies and chemistry at a reasonable cost. Processing of the titanium-45 was straightforward and simpler than in previously reported procedures. Imaging studies demonstrated that the decay properties of titanium-45 are well suited for small animal PET, justifying further investigation into possible ligand systems for this radionuclide. Acknowledgments The authors thank Bill Margenau, Pat Margenau, Dave Ficke and Todd Perkins for their technical help. Thanks also goes to Michael R. Lewis, Ph.D., and Douglas J. Rowland, Ph.D., for their scientific contributions, and Jason S. Lewis, Ph.D., for his helpful discussions. This research was

122

A.L. Va¯vere et al. / Nuclear Medicine and Biology 32 (2005) 117–122

supported by the United States Department of Energy (DEFG02-87ER60212, M.J.W.). References [1] Koepf-Maier P, Preiss F, Marx T, Klapoetke T, Koepf H. Tumor inhibition by titanocene complexes: activity against sarcoma 180. Anticancer Res 1986;6(1):33 – 7. [2] Koepf-Maier P, Wagner W, Koepf H. In vitro cell growth inhibition by metallocene dichlorides. Cancer Chemother Pharmacol 1981;5(4): 237 – 41. [3] Koepf-Maier P. Tumor inhibition by titanocene complexes: influence on xenografted human adenocarcinomas of the gastrointestinal tract. Cancer Chemother Pharmacol 1989;23(4):225 – 30. [4] Koepf-Maier P. Complexes of metals other than platinum as antitumour agents. Eur J Clin Pharmacol 1994;47(1):1 – 16. [5] Christodoulou CV, Ferry DR, Fyfe DW, Young A, Doran J, Sheehan TMT, et al. Phase I trial of weekly scheduling and pharmacokinetics of titanocene dichloride in patients with advanced cancer. J Clin Onc 1998;16(8):2761 – 9. [6] Harding MM, Mokdsi G. Antitumor metallocenes: structure-activity studies and interactions with biomolecules. Curr Med Chem 2000; 7(12):1289 – 303. [7] Toney JH, Marks TJ. Hydrolysis chemistry of the metallocene dichlorides M(D5-C5H5)2Cl2, M = titanium, vanadium, or zirconium. Aqueous kinetics, equilibria, and mechanistic implications for a new class of antitumor agents. J Am Chem Soc 1985;107(4):947 – 53. [8] Guo M, Sun H, McArdle HJ, Gambling L, Sadler PJ. Ti(IV) uptake and release by human serum transferrin and recognition of Ti(IV)transferrin by cancer cells: understanding the mechanism of action of the anticancer drug titanocene dichloride. Biochemistry 2000; 39(33):10023 – 33. [9] Lederer CM, Shirley VS. Table of isotopes. 7th ed. New York (NY)7 Wiley; 1978. [10] Laforest R, Rowland DJ, Welch MJ. MicroPET imaging with nonconventional isotopes. IEEE Trans Nucl Sci 2002;49(5 Pt 1): 2119 – 26.

[11] Merrill JC, Lambrecht RM, Wolf AP. Cyclotron isotopes and radiopharmaceuticals. XXIV. Titanium-45. Int J Appl Radiat Isot 1978;29(2):115 – 6. [12] Ishiwata K, Ido T, Monma M, Murakami M, Kameyama M, Fukuda H, et al. Preparation of 45Ti-labeled compounds and their medical application. J Label Compd Radiopharm 1982;19:11 – 2. [13] Kawamura M, Ido T, Ishiwata K, Inoue K, Kimura A, Matsuda K, et al. Metabolism of 45Ti-labeled compounds: effect of ascorbic acid. J Label Compd Radiopharm 1986;23:1360 – 2. [14] Ishiwata K, Ido T, Monma M, Murakami M, Fukuda H, Kameyama M, et al. Potential radiopharmaceuticals labeled with titanium-45. Appl Radiat Isot 1991;42(8):707 – 12. [15] McCarthy DW, Shefer RE, Klinkowstein RE, Bass LA, Margeneau CS, Cutler CS, et al. Efficient production of high specific activity 64 Cu using a biomedical cyclotron. Nucl Med Biol 1997;24(1): 35 – 43. [16] Lewis MR, Reichert DE, Laforest R, Margenau WH, Shefer RE, Klinkowstein RE, et al. Production and purification of gallium-66 for preparation of tumor-targeting radiopharmaceuticals. Nucl Med Biol 2002;29(6):701 – 6. [17] Knoess C, Siegel S, Smith A, Newport D, Richerzhagen N, Winkeler A, et al. Performance evaluation of the microPET R4 PET scanner for rodents. Eur J Nucl Med Mol Imaging 2003;30(5):737 – 47. [18] Dell Jr GF, Ploughe WD, Hausman HJ. Total reaction cross sections in the mass range 45 to 65. Nucl Phys 1965;64(3):513 – 23. [19] Levkovskij VN. Activation cross section nuclides of average masses (A= 40–100) by protons and alpha-particles with average energies (E = 10–50 MeV). Activation cross-sections by protons and alphas. Moscow; 1991. Accessed: http://bnlnd2.dne.bnl.gov/. [20] Derenzo SE, Budinger TF, Huesman RH, Cahoon JL, Vuletich T. Imaging properties of a positron tomograph with 280 BGO crystals. IEEE Trans Nucl Sci 1981;NS28(1):81 – 9. [21] Dean JA. Lange’s handbook of chemistry. 14th ed. New York7 McGraw-Hill; 1992. [22] Clark RJH. The chemistry of titanium and vanadium, an introduction to the chemistry of the early transition elements (topics in inorganic and general chemistry). New York7 Elsevier; 1968.