Development and long-term evaluation of a new 68Ge/68Ga generator based on nano-SnO2 for PET imaging

Radionuclide generator systems can routinely provide radionuclides on demand such as 68Ga produced by a 68Ge/68Ga generator without the availability of an on-site accelerator or a research reactor. Thus, in this work nano-SnO2 was used to develop a new 68Ge/68Ga generator which was evaluated over a period of 17 months and 305 elution cycles. The elution yield was 91.1 ± 1.8% in the first 7 mL (1 M HCl as eluent) when the generator was new and then it decreased with time and use to 73.8 ± 1.9%. Around 80% of the elutable 68Ga activity was obtained in 1 mL and the 68Ge content in the eluate did not exceed 1 × 10–4% over the investigation period when it was eluted regularly. The described generator provided adequate results for radiolabelling of DOTA-TOC with direct use of eluate. In addition, [68Ga]Ga-DOTA-TOC was tested satisfactorily for in vivo tumor detection by microPET/CT imaging in a lung cancer mouse model.

68 Ga-based radiopharmaceuticals are very relevant to Positron Emission Tomography (PET) because this radionuclide has decay characteristics suitable for PET imaging 1 . 68 Ga disintegrates to 68 Zn by positron emission (88.88%) with a maximum energy of 1899.1 keV and partially by electron capture (11.11%), and has a physical half-life (T 1/2 ) of 67.845 ± 0.018 min (mean ± variance) 1 , pharmacokinetically compatible with a number of molecules with a high clearance such as those of low molecular weight, antibody fragments and peptides. Due to the fact the 68 Ga can form stable complexes with many ligands containing donor atoms such as nitrogen and oxygen means that a great variety of chelating agents can be used for complexation. 68 Ga can also form macromolecules useful in the diagnosis, treatment and follow-up of patient response to chemo-and radiotherapy 2,3 .
Gallium-68 is the product of the radioactive decay of 68 Ge and can be obtained by using a 68 Ge/ 68 Ga generator. Germanium-68 has a T 1/2 = 270.95 ± 0.26 days 4 , and this permits the manufacture of potentially inexpensive and long-lived generator systems 5 . This is an advantage over usual PET radionuclides produced via proton irradiation such as 18 F, 11 C, 13 N, 15 O, 64 Cu and 124 I. An additional benefit of using 68 Ga is that it can be separated easily on demand and its availability in highly-specific activity, no-carrier-added form 6,7 . Furthermore, the higher resolution of PET when compared to Single Photon Emission Computed Tomography (SPECT) could result someday in 99m Tc or 111 In-based radiopharmaceuticals being replaced by those based on 68 Ga.
The most common commercially available 68 Ge/ 68 Ga radionuclide generators nowadays are discussed here. Despite remarkable advancements, the content of 68 Ge in the eluate (breakthrough) and metallic impurity presence in generator eluate appear to be the greatest impediments to direct-labelling of 68 Ga in the manufacture of 68 Ga-based radiopharmaceuticals. Potential metal ions from generator eluate may challenge 68 Ga 3+ for coordinative labelling of radiopharmaceutical precursors and therefore produce lower 68 Ga incorporation or a reduction of specific radioactivity in the ultimate 68 Ga-labelled radiopharmaceutical 3 . The European Pharmacopoeia 8 specifies that gallium ( 68 Ga) chloride solution shall not contain more than 10 µg of Fe and Zn per GBq of 68 Ga in eluate that is used for radiolabelling and that no more than 0.001% of 68 Ge should be present in the 68 Ga eluate. Table 1 gives an outline of the characteristics of major commercial 68 Ge/ 68 Ga generators currently available on the market. Two of the most commonly-used, commercially-available 68  pletely adsorbed in the nano-SnO 2 column and the 68 Ge/ 68 Ga generator was evaluated for preclinical research applications by examination of its performance over an extended period of 17 months (305 elutions). The generator was designed to operate either in the "dry" or "wet" operation mode (see "Materials and methods"). Over 70% of the 68 Ga was eluted with 7 mL of 0.5-1 M HCl, as shown in Fig. 1. The elution yield of 68 Ga was higher when eluted with 1 M HCl (91.1 ± 1.8%; n = 50), compared to elution with 0.5 M HCl (70.1 ± 3.5%; n = 68) when the generator was operated in the "wet" mode at 1 mL/min (Supplementary Table S1 online). Therefore, the generator was eluted using 1 M HCl from elution number 69 onward. This is in good agreement with previous results that described a 68 Ge/ 68 Ga generator based on SnO 2 where the activities of 68 Ga in the eluate increased when HCl concentration was raised from 0.5 to 10 M 23 . While elution with 2 M HCl could result in greater 68 Ga activity in the eluate, as we have reported with an in-house built pilot 68 Ge/ 68 Ga generator based on nano-SnO 2 22 , we preferred to elute the generator with lower acidity, at a maximum of 1 M, in order to preserve generator column integrity. 68 Ga (1 M HCl, 1 mL/min) elution yield gradually decreased over time, ranging from 91.1 ± 1.8% (n = 50) to 73.8 ± 1.9% (n = 27) (Supplementary Table S1 online); these values are typical of 68 Ge/ 68 Ga generators based on inorganic oxides 7,[9][10][11][12]24 . As was seen with the CeO-PAN-based 68 Ge/ 68 Ga generator, the decrease in elution yield over time was lower than was observed with commercial generators, in which the elution yield decreased to ~ 50% of its initial value 20 . Beyond the amount of eluted 68 Ga, another great advantage of our generator is that its design permits operation in either "dry" or "wet" mode; no differences were found when the generator was operated in either mode (around 75-85%), thus confirming our previous findings 22 . An advantage of using HCl is that the environment it creates is hostile to microorganisms. This phenomenon has been reported when studying the survival of microorganisms such as fungal spores and acid-resistant bacteria in a generator column eluted with 0.1 M HCl. In this environment no viable microorganisms could be detected even over a prolonged testing period of several days. Risks associated with incidental microbial contaminations during the lifetime of such a generator were very low, as the conditions used in this type of generator were highly unfavourable for survival or growth of microorganisms 24 .
Germanium-68 breakthrough was expressed as a percentage of 68 Ge activity in the eluate with respect to 68 Ge loaded at calibration time. 68 Ge breakthrough from the generator was negligibly low (< 10 -4 %) and remained unchanged over the period studied ( Fig. 1 and Supplementary Table S1 online). The generator's quality in terms of 68 Ge breakthrough was high, a result that has been confirmed in visualization by PET imaging of 68 Ge adsorbed into the nano-SnO 2 -based sorbent. The PET image of the 68 Ge/ 68 Ga generator column after 305 elutions showed highest activity in the upper two-thirds of the column, i.e., no significant spreading and drifting of 68 Ge zone in the column was observed after 17 months of use (Fig. 2). A slight increase in 68 Ge breakthrough was observed on occasion when the generator was not eluted for more than 3 days, which is a phenomenon that had been observed in our previous study 22 . Nevertheless, after a few elutions, this always returned to values below 10 -4 %. For this reason daily elutions are recommended to keep the elution's 68 Ge content constant. We obtained the best results with the present nano-crystalline calcined SnO 2 -based 68 Ge/ 68 Ga generator (740 MBq of 68 Ge) compared with our nano-SnO 2 -based in-house built pilot 68 Ge/ 68 Ga generator (1.85 MBq of 68 Ge) which we had previously reported 22 . This can be explained by a greater amount of stationary phase used to prepare the generator with higher activity. The breakthrough obtained from our generator was considerably lower than with commercial 68 Ge/ 68 Ga generators 7,[9][10][11][12][13][14]24 . These generators commercially supplied using SnO 2 are based on the use of bulk tin oxide as sorbent and suffer a decrease in the elution yield of 68 Ga and an increase of the breakthrough of 68 Ge over a prolonged period of time 13 . An alternative is to use mesoporous metal oxides as sorbents for 68 Ge/ 68 Ga www.nature.com/scientificreports/ generators since they have demonstrated good performance and proven efficiency 25 . The average pore radius of our nano-SnO 2 -based 68 Ge/ 68 Ga generator was 14.8 nm indicating that the nano-crystalline SnO 2 sorbent was mesoporous/macroporous. The mean crystal size was found to be 28.1 ± 0.5 nm; moreover, the particles of calcined SnO 2 had a spherical shape with a high degree of agglomeration among fine particles 22 . The use of micron sized agglomerated particles is essential in the preparation of radionuclide generators since they decrease chemical impurities in the eluate 25 . Figure 3 shows the elution profiles of both radionuclides 68 Ga and 68 Ge, collecting fractions of 0.5 mL at a flow rate of 1 mL/min with 7 mL of 1 M HCl. The majority of 68 Ga activity was eluted in fractions 2 and 3 when it was operated in "dry" mode (81.79 ± 1.00%; n = 3) and in fractions 4 and 5 when the generator was operated in "wet" mode (82.05 ± 1.12%; n = 3). While the percentage of the eluted 68 Ge activity was very similar in all the fractions eluted in "wet" mode (5.01 ± 0.08%), a slight increase in the percentage of activity eluted in the first 1 mL (12.3 ± 3.4% and 6.5 ± 0.3% in the first and second fraction, respectively) was observed when the generator was operated in "dry" mode ( Fig. 3). During the period studied the elution profile of both radionuclides remained unchanged. Therefore, if the generator is going to be operated in "wet" mode and its eluate will be used directly for 68 Ga labelling procedures, the first 1.5 mL can be discarded and the subsequent 1 mL (containing the highest concentration of pure 68 Ga) collected and buffered to an appropriate pH for radiolabelling. The elution profile  6 8 Ga -wet 6 8 Ga -dry 6 8 Ge -dry 6 8 Ge -wet www.nature.com/scientificreports/ obtained in "wet" mode was comparable to that reported in a detailed study of a commercial SnO 2 generator 26 and also with other commercial 68 Ge/ 68 Ga generators based on SnO 2 , TiO 2 , SiO 2 and CeO 2 -PAN 13 . Analysis of the selected decayed samples of 68 Ga by Coupled Plasma-Mass Spectrometry (ICP-MS) revealed that levels of Ag, As, Ba, Bi, Ca, Cd, Co, Cr, Cu, Fe, In, Mg, Mn, Ni, Pb, Sb, Sn and Ti were always < 1 ppm, as shown in Table 2. The most frequently present impurity was Zn, primarily due to 68 Ga decay and potential contamination in the preparation of 0.5-1 M HCl used to eluate the generator. In the 68 Ge/ 68 Ga generator column, the amount of Zn exceeds 10 times that of 68 Ga at the time of secular equilibrium 27 , consequently, when the generator is eluted, if all the Zn in the column was eluted in 1 mL of 1 M HCl, the level of Zn would be negligibly low (< 0.1 ppm). "Dry" mode use of the 68 Ge/ 68 Ga generator resulted in Zn levels that were lower than when the "wet" mode was used ( Table 2). When the generator was operated in "dry" mode, concentration Zn was below 0.5 ppm in half of the samples and in the other half, the mean concentration was 6.4 ± 7.9 ppm. Considerable care must therefore be taken during preparation of HCl solutions and samples for ICP-MS (zinc is a common contaminant in laboratories 28 ). With regards to other metals, Fe and Cu ions are the most significant competitors with 68 Ga 3+ in the complexation process with bifunctional chelators 29 . The concentrations determined for both ions were extremely low, always under 100 ppb ( Table 2). A common impurity expected in the 68 Ga eluate of commercial SnO 2 -based generators is the Sn ion due to the dissolution of the column matrix. It is of interest that Sn concentration was low, even under 100 ppb when the generator is operated in "dry" mode. Increased concentration of trace metals was not observed in the first elution relative to the second day or after (data not shown), which is something that has been described by other authors 10 . The elution was directly evaluated for radiolabelling of DOTA-TOC without performing any additional purification process due to the relatively low concentration of trace metals.
Radiolabelling. 68 Table S2, online). In addition, we prepared other 68 Ga-complexes www.nature.com/scientificreports/ with high yields using ligands such as NODAGA-(RGD) 2 , PSMA-11 and iron-oxide nanoparticles (data not shown). Although some authors have replaced HEPES by sodium acetate as radiolabelling buffer (since this is an easily available and pharmacologically harmless buffer 30) , we decided to use HEPES because it increases the radiolabelling yield and, although HEPES is currently not approved for human use, we believe that it is the most appropriate buffer for use in the preparation of 68 Ga-DOTA-peptides for small animal PET imaging. The eluate of 68 Ge/ 68 Ga generator is not chemically pure so non-radioactive metallic impurities may compete with 68 Ga for coordinative labelling of radiopharmaceuticals precursors 31 . The metal ion content of the 68 Ge/ 68 Ga generator eluate depends on elution frequency. Therefore, a preventive elution 4 h prior to the synthesis has been recommended since the interfering metal ion concentration can be kept to a minimum 10 . We saw no differences between the first elution of the generator each day and an elution 3-4 h later with respect to labelling yield, which perhaps is due to the low amount of metal ion traces present in the eluate. Additionally, using an excess of precursor in the synthesis procedure could contribute to this situation.
Representative HPLC-radiochromatogram before and after purification from the C-18 Sep-Pak (Fig. 4a,b) showed a peak corresponding to [ 68 Ga]Ga-DOTA-TOC with retention time of 9.94 min. Free 68 Ga showed a retention time of 2.78 min (Fig. 4a). After Sep Pak purification, free 68 Ga was removed and radiochemical purity (RQP) of the final product was determined to be ≥ 97%. An unresolved front shoulder before the mean peak could be observed in both chromatograms; this poorly separated shoulder was estimated to be < 3% of total radioactivity regardless of the amount of starting activity. These radioimpurities are partially attributable to the radiolytic oxidation of [ 68 Ga]Ga-DOTA-TOC, which is consistent with previous results 26 . The addition of radical scavengers such as ascorbic acid could significantly suppress by-product formation 32 . pet imaging with [ 68 Ga]Ga-DotA-toc in a preclinical mouse model of lung cancer. We used lung cancer mouse models generated in the laboratory to determine the suitability of the [ 68 Ga]Ga-DOTA-TOC obtained for detection of tumors in vivo by PET imaging. We have used a high-grade neuroendocrine lung cancer animal model rendering Large-Cell Neuroendocrine Carcinoma (LCNEC) 33,34 . LCNEC is an aggressive, poorly differentiated pulmonary neuroendocrine tumor which expresses a high density of surface somatostatin receptors (SSTR). This allows imaging with radiolabelled somatostatin analogue DOTA-TOC 35 .
Micro-PET/CT [ 68 Ga]Ga-DOTA-TOC showed increased uptake in some areas of the lungs of the cancer developing mouse models (Fig. 5a). No signal, or background signal, was detected in healthy control mice (not shown). Histological analyses allowed accurate lung cancer diagnosis which revealed Large-Cell Neuroendocrine Carcinoma in the lungs of the tested mice (n = 4) (Fig. 5b). Expression of Somatostatin Receptor 2 (SSTR2) in the tumors (Fig. 5c) was confirmed by immunohistochemistry. These studies demonstrate that early detection of these tumors is feasible due to their avidity for 68 Ga-DOTATOC that allows follow-up and monitoring changes after therapy. Thus, small animal in vivo PET imaging tumor detection using the [ 68 Ga]Ga-DOTA-TOC is a powerful tool in preclinical evaluation of cancer development with potential translational and therapeutic applications.
conclusions Nano-crystalline SnO 2 provided column support for a 68 Ge/ 68 Ga generator system which was prepared and evaluated over a period of more than 1 year of use and 300 elutions. The generator produced readily and reproducibly carrier-free 68 Ga by elution with 1 M HCl solution. The elution yield of 68 Ga was > 90% in the first 7 ml when the generator was new and subsequently decreased with intensive use to around 70%. 68 Ge content in the eluate did not exceed 1 × 10 -4 %. Using either the "wet" or "dry" elution mode of the 68 Ge/ 68 Ga generator did not result in any noticeable differences in elution yield, therefore, the use of "dry" elution mode may be preferable for long-term applications in order to prevent tin oxide from being slightly attacked by the HCl used as eluent 36 . Furthermore, 68 Ga direct elution from the generator, without the inclusion of a post-elution purification procedure, was successfully used for radiolabelling DOTA-TOC. The radiotracer [ 68 Ga]Ga-DOTA-TOC showed a remarkable sensitivity for the in vivo detection of large-cell neuroendocrine tumors in a lung cancer mouse model. For the purpose of radioprotection, the column was kept in an easily-transportable lead shielded cylindrical shape (250 × 10 × 55 mm; height × inner radius × external radius) and all operations were carried out in a closed system using connecting Teflon tubes. A sterile polyethylene 10 mL syringe was attached to the inlet tube via a Luer-lock connection to elute 68 Ga from the generator with HCl (Fig. 2a). Excess of tin dioxide, synthesized on demand by Keeling & Walker (UK) and previously characterized by our research group 22 , was calcined at 900 °C for 3 h in an oven and cooled later to room temperature. Next, it was sifted through a 100-150 µm sieve in a stainless steel frame. A borosilicate glass column was packed with 4.5 g of nano-SnO 2 and preconditioned with 50 mL of 0.5 M HCl to remove fine particles. Subsequently the column was loaded with 740 MBq of 68 Ge in no-carrier-added form (obtained from JSC Isotope, Russia) and washed with 50 mL of 1 M HCl solution in order to remove trace levels of non-adsorbed 68 Ge. One day after loading 68 Ge, enough time for in-growth of the 68 Ga activity, the generator was ready to use.
Gallium-68 was eluted from the generator using 7 mL of hydrochloric acid (0.5-1 M HCl) prepared from a solution of 34% hydrochloric acid, Ultrex II ultrapure reagent (J. T. Baker) at a constant flow rate of 1 or 1.5 mL/ min provided by an infusion pump. The generator was designed to operate either in the "dry" or "wet" mode of operation. In the "dry" mode, after each elution the rest of HCl was removed by passing air through the column; it often happens that some HCl remains in the column despite drying. In the "wet" mode, after each elution the column remains filled with HCl till the next elution. The performance of the generator was evaluated for 17 months, eluting it regularly at intervals of once, twice or three times per day (more than 300 elutions during the period of study). The elution profile was studied by collecting the eluates in fractions of 0.5 mL and determining their activity. A dose calibrator model VDC-405 (Veenstra Instruments) was used for the determination of 68 Ga activity after elution. The elution yield was expressed as the percentage of the eluted 68 Ga activity in relation to the theoretical activity of 68 Ga in the generator column after decay correction for the elapsed time. As the half-life of the 68 Ge is much greater than the 68 Ga, the theoretical activity of 68 Ga was calculated by using the equation corresponding to secular radiochemical equilibrium: www.nature.com/scientificreports/ where A p and A d are the activity of 68 Ge and 68 Ga respectively and λ d (0.0102 min −1 ) 1 is the radioactive decay constant of 68 Ga. The activity of 68 Ge in aliquots of eluates was analyzed by gamma-ray spectrometry using a calibrated coaxial High Purity Germanium (HPGe) detector model GEM 13,200 (Ortec) coupled to multichannel gamma-ray analyzer DSA1000 (Canberra). The generator eluates were analyzed after at least 48 h in order to allow the 68 Ga to decay to a level that permits the detection of 68 Ge since it can be detected indirectly as a decay product, 68 Ga. All samples, which contained 1 mL, were measured using the same conditions: constant geometry using 1.5 mL conical bottom tubes made from polypropylene, placed at a distance of 5 cm from the detector and 6 h as minimum measurement time. The activity of each sample was calculated by quantifying the area of 511 keV peak in the spectrum obtained, which corresponds to radioactive decay of 68 Ga. The Minimum Detectable Activity (MDA) for 68 Ga was 1 Bq. Finally, the 68 Ge activity was calculated by direct comparison with the results obtained from the measurement of a standard source of 68 Ge, performed by Ionizing Radiations Metrology Laboratory belonging to the CIEMAT, under the same measurement conditions previously mentioned. The generator column was visually checked additionally by small animal PET/CT SuperArgus (SEDECAL, Spain) with the purpose of studying the possible spread of 68 Ge zone along the column due to the number of elutions and the time of use of the generator.
The presence of potential chemical impurities (in the form of Ag, Al, As, Ba, Bi, Ca, Cd, Co, Cr, Cu, Fe, In, Mg, Mn, Ni, Pb, Sb, Sn, Ti and Zn) in the 68 Ga eluate was determined in three groups according to the elution number: 24-61, 75-124 and 263-275. The trace level detection of the metal ion contamination in decayed samples were carried out by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) using a quadrupole single collector iCAP Q (Thermo Scientific) equipped with a collision cell. The water used was Milli-Q purity (Millipore) and the acids were purified by Sub-Boiling distillation using a duoPUR system (Milestone). The non-disposable materials used were previously washed with 5% HNO 3 and distilled water. Quantification was carried out by external calibration and internal standard, preparing the standards by successive dilution of a certified standard (Alfa Aesar and Inorganic Ventures) to 2-5% (v/v) using 65% HNO 3 (Sigma-Aldrich) for each of the elements.
Radiolabelling. Labelling properties of the 68 Ge/ 68 Ga generator were tested by a method based on the radiolabelling of DOTA-TOC (DOTA-[Tyr3]-octreotide), provided by ABX (Radeberg, Germany). This peptide, a DOTA derivatized somatostatin analog, shows high affinity to the SSTR2 subtype of somatostatin receptor expressing tumors and binds the trivalent Ga 3+ with a high kinetic and thermodynamic stability. It is widely used in clinical routines for PET imaging of neuroendocrine tumors 37 . The procedure used to label, purify and determine the radiochemical purity of [ 68 Ga]Ga-DOTA-TOC was essentially the same as that used by our group 33 with some adjustments. A solution of 210 ± 5 mg HEPES (Sigma-Aldrich) dissolved in 1 mL of deionized water was transferred into a reaction vial which contained the peptide DOTA-TOC previously dissolved in water. The 68 Ge/ 68 Ga-generator was eluted with 7 ml of 1 M HCl solution (prepared from 34% HCl, Ultrex II ultrapure reagent, J. T. Baker) and 1 mL of the eluate was transferred to the reaction vial. The final pH of the reaction mixtu re was 3.5-4, measured by indicator paper strips (Merck). After very careful shaking, the mixture was heated for 5 min at 90 ± 5 ºC using a microwave with monomodal radiation and after cooled to room temperature with nitrogen.
The crude product was purified by solid-phase extraction cartridge Sep Pak light C18 cartridge (Waters) previously conditioned and equilibrated with 4 mL of pure ethanol and 4 mL of deionized water. The mixture was transferred onto Sep Pak where [ 68 Ga]Ga-DOTA-TOC and colloidal 68 Ga were retained, whereas free 68 Ga passed through the cartridge. Next, the cartridge was rinsed with 4 mL of deionized water and the activity on the Sep Pak cartridge was reversely recovered with 0.5 mL ethanol 96% (v/v) (Scharlab); during this procedure all colloidal 68 Ga was retained on the Sep Pak cartridge. Finally, ethanol was evaporated to dryness using a Speed Vac (Savant) and [ 68 Ga]Ga-DOTA-TOC was dissolved with a 0.9% NaCl solution (Braun). The radiolabelling yield was calculated by comparing the fraction of 68 Ga collected in ethanol with the activity in the initial radiolabelling reaction using decay-corrected radioactivity values.
The radiochemical purity of [ 68 Ga]Ga-DOTA-TOC was determined by radio-High Performance Liquid Chromatography (radio-HPLC) (Jasco) equipped with a photo diode array detector MD-4015 (Jasco), a radioactivity detector LB 507A (Berthold) and the measurements performed under the following conditions. Column: Jupiter Proteo 90 Å column (Phenomenex), 4 µm, 250 × 4.6 mm. Eluents: 0.1% trifluoroacetic acid (TFA) in 5% acetonitrile (solvent A) and 0.1% TFA in 95% acetonitrile (solvent B). Gradient: 0-2 min (5% B), 2-17 min (5-100% B), 17-20 min (100% B), 20-22 min (100-5% B) and 22-24 min (5% B). Flow rate: 1 mL/min. Two aliquots, before and after purification step by Sep Pak, were taken to determine radiochemical purity. TFA was purchased from Sigma-Aldrich and Acetonitrile Ultra Gradient HPLC grade from J. T. Baker. preclinical model for pulmonary neuroendocrine tumors. All animal experiments were approved by the Animal Ethical Committee of the Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT) and conducted in compliance with institutional and national guidelines. A mouse model of high-grade neuroendocrine lung cancer (LCNEC) based on the ablation of four tumor suppressor genes (Rb, Rbl1,pTEN and Trp53) was used to perform imaging studies 33, 34, 38 . pet/ct imaging. PET studies were performed in an Argus PET/CT (SEDECAL) scanner. The procedure used was as described by our group 33  www.nature.com/scientificreports/ were performed at 90 min after intravenous injection of [ 68 Ga]Ga-DOTA-TOC (2)(3)(4)(5) in mice anesthetized by inhalation of 2-2.5% of 2-2.5% isoflurane (Esteve) in 100% oxygen (ALPHAGAZ™, Air Liquide) at a flow rate of 5% oxygen using a Fluovac System (Harvard Bioscience). PET images were corrected for random events and scatter with and without attenuation correction and reconstructed using the 2D-OSEM (Ordered Subset Expectation Maximization) algorithm (16 subsets and 2 iterations). For image-reading purposes PET images were fused with the corresponding CT images, which were used as a reference for image co-registration. Images were analyzed using the image analysis software AMIDE, version 1.0.4 (https ://amide .sourc eforg e.net).
Histology and Immunohistochemistry. Histology and Immunohistochemistry tests were performed to confirm the diagnosis of the tumor and the expression and localization of SSTR2. Lungs were perfused at necropsy with 4% formaldehyde, fixed and embedded in paraffin wax sections (5 μm) were stained with hematoxylin and eosin (H&E) for histological analysis or processed for immunohistochemistry, essentially as in previously described standard protocols 39 . For expression of SSTR2, a high temperature antigen unmasking technique consisting of 15-min microwaving of slides in 0.01 M citrate buffer (Antigen Retrieval Buffer, Abcam) was used after deparaffinization to enhance the staining. Sections were then incubated with 5% horse serum (Gibco) for 30 min to block the Fc receptor in tissue, and then washed three times with sterile PBS (Gibco, pH 7.5) prior to incubation with the anti-SSTR2 primary antibody (Sigma-Aldrich, diluted 1:100). Peroxidase-conjugated secondary antibody supplied by the manufacturer was used at 1:300 dilution and visualized using diaminobenzidine as a substrate (DAB kit Vector). Control slides were obtained by replacing primary antibodies with PBS (data not shown).

Statistical analyses.
Continuous measurements are presented as mean ± standard deviation (μ ± σ), unless otherwise stated. Mean values were compared using one-way ANOVA followed by post hoc testing or custom contrasts. The difference was considered statistically significant when p value was < 0.05. Statistical analyses were performed using SPSS in 14.0.

Data availability
The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.