F-18 labelled PSMA-1007: biodistribution, radiation dosimetry and histopathological validation of tumor lesions in prostate cancer patients
The prostate-specific membrane antigen (PSMA) targeted positron-emitting-tomography (PET) tracer 68Ga-PSMA-11 shows great promise in the detection of prostate cancer. However, 68Ga has several shortcomings as a radiolabel including short half-life and non-ideal energies, and this has motivated consideration of 18F-labelled analogs. 18F-PSMA-1007 was selected among several 18F-PSMA-ligand candidate compounds because it demonstrated high labelling yields, outstanding tumor uptake and fast, non-urinary background clearance. Here, we describe the properties of 18F-PSMA-1007 in human volunteers and patients.
Radiation dosimetry of 18F-PSMA-1007 was determined in three healthy volunteers who underwent whole-body PET-scans and concomitant blood and urine sampling. Following this, ten patients with high-risk prostate cancer underwent 18F-PSMA-1007 PET/CT (1 h and 3 h p.i.) and normal organ biodistribution and tumor uptakes were examined. Eight patients underwent prostatectomy with extended pelvic lymphadenectomy. Uptake in intra-prostatic lesions and lymph node metastases were correlated with final histopathology, including PSMA immunostaining.
With an effective dose of approximately 4.4–5.5 mSv per 200–250 MBq examination, 18F-PSMA-1007 behaves similar to other PSMA-PET agents as well as to other 18F-labelled PET-tracers. In comparison to other PSMA-targeting PET-tracers, 18F-PSMA-1007 has reduced urinary clearance enabling excellent assessment of the prostate. Similar to 18F-DCFPyL and with slightly slower clearance kinetics than PSMA-11, favorable tumor-to-background ratios are observed 2–3 h after injection. In eight patients, diagnostic findings were successfully validated by histopathology. 18F-PSMA-1007 PET/CT detected 18 of 19 lymph node metastases in the pelvis, including nodes as small as 1 mm in diameter.
18F-PSMA-1007 performs at least comparably to 68Ga-PSMA-11, but its longer half-life combined with its superior energy characteristics and non-urinary excretion overcomes some practical limitations of 68Ga-labelled PSMA-targeted tracers.
Keywords18F-PSMA F-18-PSMA PSMA-1007 PET/CT Positron emission tomography
The introduction of the 68Ga-labelled prostate-specific membrane antigen (PSMA) targeted positron-emitting-tomography/computed-tomography (PET/CT) tracer Glu-urea-Lys(Ahx)-HBED-CC (PSMA-11) has proven highly sensitive for the detection of disseminated prostate cancer (PCa). In two studies (involving 319 patients and 248 patients, respectively) sites of biochemical recurrence (BCR) were localized in 90 % of patients including those with modest elevations in prostate specific antigen (PSA) [1, 2]. PSMA-11 PET/CT appears superior in sensitivity to other PET agents such as Choline-PET/CT [3, 4]. This high sensitivity could have significant clinical implications for modifications of treatment at various stages of prostate cancer ranging from initial diagnosis to treatment monitoring of castration resistant metastases. For instance, 68Ga-PSMA-11 PET/CT had a significant impact on radiotherapy planning, resulting in meaningful changes in treatment planning in over 50 % of patients [5, 6]. Although extensively studied in the recurrence and metastatic setting, there has been relatively little attention paid to the use of PSMA-PET in initial staging [7, 8, 9, 10]. One challenge of 68Ga-PSMA-11 imaging is that the agent is rapidly excreted via the urinary tract resulting in intense accumulation in the bladder, thus, obscuring the prostate. This has resulted in the development of other agents with slower urinary excretion. One such candidate, 99mTc-MIP-1404, has been chosen for evaluation in phase-2 (NCT01667536) and phase-3 (NCT02615067) studies in North America for primary staging of PCa . However, 99mTc is a single photon emitter and thus, has neither the sensitivity nor can gain the spatial resolution of PET-based agents. Another challenge for 68Ga-PSMA-11 PET/CT is the limited availability of the 68Ga via local radionuclide generators. Each generator provides only one or two elutions per day and is not only a substantial upfront investment but requires separate syntheses at different times of the day in a local radiopharmacy. Compared to 18F (0.65 MeV), the positron energy of 68Ga is higher (1.90 MeV), reducing the theoretical maximum spatial resolution . Finally, the short half-life of 68Ga relative to 18F, (68 vs. 110 min) limits the ability to produce agents in a central facility and ship them to distributed imaging centers.
Material and methods
Synthesis and quality control of 18F-PSMA-1007
The non-radioactive reference compound 19F-PSMA-1007 and the PSMA-1007 precursor were synthesized using solid phase chemistry . 18F-PSMA-1007 was produced on an automated synthesis module (Trasis AllInOne) in a two-step synthesis using the prosthetic group 6-[18F]F-Py-TFP  for coupling the PSMA-1007 precursor, and this was purified by semi-preparative HPLC. The detailed description of the radiolabelling process was published elsewhere . A direct radiofluorination method is being developed. Radio-HPLC was performed to determine the chemical identity and the chemical and radiochemical purity of 18F-PSMA-1007. The chemical structure of 18F-PSMA-1007 is presented in Fig. 1. Residual solvents were determined by gas chromatography. Radionuclide purity was controlled by half-life measurement. Integrity of the sterile filter after filtration was assessed using the bubble-point test. The product solution was tested for sterility, bacterial endotoxins (LAL-test), pH, colorlessness and particles.
All imaging was performed on a Biograph mCT Flow scanner (Siemens, Erlangen, Germany). PET was acquired in 3-D mode (matrix 200 × 200) using FlowMotion (Siemens). The emission data was corrected for randoms, scatter and decay. Reconstruction was performed with an ordered subset expectation maximization (OSEM) algorithm with two iterations/21 subsets and Gauss-filtered to a transaxial resolution of 5 mm at full-width at half-maximum (FWHM); Attenuation correction was performed using the unenhanced low-dose CT data. The CT-scans were reconstructed to a slice thickness of 5 mm, increment of 3–4 mm, soft tissue reconstruction kernel (B30), using CareDose (Siemens).
Volunteers and patients
To determine initial dosimetry, three healthy volunteers (normal PSA) underwent 18F-PSMA-1007 PET/CT with scans obtained at multiple time points up to 6 h p.i. The healthy subjects were imaged in three blocks. Block 1 began at PET-1 (start 5 min p.i.) and extended to PET-7 (ending 140 min p.i.), block 2 began at PET-8 (start 180 min p.i.) and extended to PET-9 (240–270 min p.i.) and block 3 began at PET-10 (440–480 min p.i.). A non-enhanced low-dose CT (estimate 1.4 mSv, respectively) for attenuation correction was performed at the beginning of each block, followed by serial emission scans without moving the volunteers in between.
Initial PSA (ng/ml)
PCa SUVmax (1 h p.i.)
PCa SUVmax (3 h p.i.)
pT3b, pN1 (4/41), L1, V0, Pn1
cT3b, cN1, cM1
pT3b, pN1 (5/40), L1, V0, Pn1
pT3b, pN1 (4/43), L0, V0, Pn1
pT3b, pN1 (3/48 LK), L1, V0, Pn1
pT3a, pN1 (3/57), L1, V1, Pn1
pT3a, pN0 (0/21), L0, V0, Pn1
pT3a, pN0 (0/32), L0, V0, Pn0
cT3, cN1, cM1
pT3a, pN0 (0/27), L0, V0, Pn1
The dosimetry analysis was performed using the QDOSE dosimetry software suite (ABX-CRO, Germany). Automatic rigid co-registration and, if necessary, additional manual correction, were performed for all PET and CT data-sets. Kidneys, liver, spleen, whole heart, upper and lower large intestine, parotid glands, submandibular glands and urinary bladder were segmented into volumes of interest (VOIs) using a percentage of maximum threshold between 20 and 30 % and the corresponding CT as guidance. Afterwards the time activity curves (TACs) were calculated for all organs. The TACs for red marrow were derived from venous blood samples, and the red marrow dose was calculated as described previously [22, 23]. The urinary bladder TAC was a combination of estimated activity in the urinary bladder VOI in PET and measured activity of voided urine. Curve fitting was applied to all TACs according to the software. Kidneys, salivary glands, upper and lower large intestine and heart were fitted with a bi-exponential function. For liver, spleen and urinary bladder content, a mono-exponential fit to the last three time points was performed.
The cumulative activity Ã between time 0 and the first measured time point was calculated assuming a linear increase from 0 to the first measured activity. The Ã between the first measured time point and the last measured time point was integrated numerically using trapezoidal approximation. The Ã from the last measured time point to infinity was integrated using the fitted function. The total body Ã was calculated based on the injected activity assuming only physical decay neglecting the voided urine. The Ã of the remainder of the body was then automatically calculated by subtracting all source organs from the total body activity. All source organ residence times were calculated by dividing the Ã by the injected activity.
Absorbed and effective dose calculations were performed using the ICRP endorsed IDAC 1.0 package  which is integrated into QDOSE. In addition, residence times of all source organs and the remainder of the body were exported as an OLINDA case file for dose calculation in OLINDA 1.1 . The absorbed doses to the salivary glands and prostate were determined using the spherical model . The organ masses for the salivary glands were taken from ICRP publication 89  with 25 g estimated weight for the parotid and 12.5 g for the submandibular gland.
During the dosimetry evaluation, urine was collected at the following intervals: 0–2 h, 2–4 h, and 4–6 h. Venous blood samples were obtained at 2, 5, 10, 15, 30, 45, 60, 80, 120, 180, 240, and 360 min post injection. After the whole blood was sampled, the remaining volume was centrifuged and serum was extracted. Serum activities were measured in a well counter and corrected for decay. Activity concentrations in blood vs. serum were compared using the individually determined (clinical routine lab) hematocrit. Total blood volume was estimated from size, weight, and hematocrit.
The tracer biodistribution in patients was quantified by SUVmean and SUVmax at 1 h and 3 h post injection. For calculation of the standardized uptake value (SUV), circular regions of interest were drawn around the area with focally increased uptake in transaxial slices and automatically adapted to a three-dimensional VOI with e.soft software (Siemens) at a 40 % isocontour. Primary tumors and lymph node metastases were evaluated separately. The normal bladder (after voiding), background (pelvic fat), blood, brain, salivary and lacrimal glands, lung, liver, spleen, pancreas, small intestine, and kidneys were evaluated with a 2–3 cm sphere placed inside the organ parenchyma.
Analyses of prostatectomy specimens were performed under the supervision of dedicated uropathologists, using International Society of Urological Pathology nomenclature . Pathologists were blinded to the PET/CT results. In addition, representative sections were stained for immunohistochemistry. To accomplish this, sections were deparaffinized in xylene and rehydrated in a graded ethanol series. Antigen retrieval was performed with a steam cooker using retrieval buffer (Target Retrieval Solution, Dako). A mouse monoclonal antibody against PSMA (clone 3E6, Dako) was used at a 1:100 dilution and incubated overnight at 4 °C. Immunodetection was performed using the Histostain-Plus detection kit (Invitrogen) according to manufacturer’s recommendations. Stained sections were scanned using a Nanozoomer 2.0-HT Scansystem (Hamamatsu Photonics) to generate digital whole slide images.
All patients and healthy volunteers tolerated the examination well. No drug-related pharmacological effects or physiologic responses occurred. All observed parameters (e.g., blood pressure, heart rate, body temperature) remained normal and unchanged during and after the examination. No patient reported subjective symptoms.
Dosimetry (OLINDA) comparison of 18F-PSMA-1007 with other PSMA-targeted tracers
Absorbed Dose (mGy/MBq)
Afshar-Oromieh et al. 
Pfob et al. 
Afshar-Oromieh et al. 
Szabo et al. 
Cho et al. 
Urinary bladder wall
Normal-organ Biodistribution and tumor uptake
Correlation of imaging and post-surgery histopathology
The primary tumors and lymph node metastases were first diagnosed on PET. Within a median of 20 days, eight patients underwent radical prostatectomy and extended pelvic lymphadenectomy (Table 1). Two patients had metastatic disease at initial diagnosis and did not undergo surgery. In total 309 pelvic lymph nodes were histologically examined, out of which 290 were negative and 19 were positive (median tumor diameter 5 mm; range 1–18 mm). 18F-PSMA-1007 PET/CT demonstrated excellent sensitivity (94.7 %). Overall, 18F-PSMA-1007 PET/CT detected 18 of 19 lymph node metastases in the pelvis including a node with a diameter as small as 1 mm. In contrast to one healthy subject who was suspicious for unspecific uptake in inflammatory lymph nodes, there were no false-positive findings in the patient cohort (specificity 100 %). Similarly, the peak intra-prostatic uptake correlated with the main tumor mass (Fig. 6). Positive PET-findings outside of the pelvis were not histopathologically validated.
In this study we evaluated the radiation dosimetry, biodistribution and preliminary efficacy of the new tracer, 18F-PSMA-1007.
Compared to other similar agents, the effective dose of 18F-PSMA-1007 PET/CT (4.4–5.5 mSv for 200–250 MBq) was comparable (Table 2). The effective organ half-life and, hence, the exposure, depends primarily on the short physical half-life of 18F rather than the biological half-life of the carrier molecule. All Glu-urea-based PSMA-targeted tracers share a similar physiological tracer distribution, thus the organ ratios differ only minimally among the various ligands. One strength of our dosimetry estimation is that we utilized a large number of ten head-to-toe whole-body scans performed up to 6 h p.i. (i.e. > 3 physical half-lives of 18F) allowing for excellent curve fitting. One source of uncertainty is the limited number of healthy subjects (n = 3). Nonetheless, since the normal organ biodistribution of 18F-PSMA-1007 has proven highly similar to other PSMA-targeted radiotracers, such as PSMA-11, PSMA-617, DCFBC and DCFPyL [13, 14, 29, 30], we consider that the results of this first dosimetry study are in keeping with prior results. The performance of dosimetry in normal volunteers is preferred to performing it in patients with large tumor masses as these could cause tumor “sink effects” distorting the normal biodistribution. This has proven to be less of a concern in PSMA ligands such as 18F-DCFPyL, in which it was shown that the dosimetry in normal organs was reliably assessed independently of the presence of tumor .
The average 1 h and 3 h SUVmax of 18F-PSMA-1007 in tumor tissue, parotid glands, liver, kidneys and spleen can be compared to data for PSMA-11 [9, 29] and PSMA-617 , because the evaluation of these 68Ga-labelled PSMA-ligands were also performed at our institution with similar imaging protocols, cross-calibrated scanners and analysis software. In contrast, 18F-DCFBC and 18F-DCFPyL have been evaluated with different scanners and imaging protocols which makes the comparison less accurate. However, even the comparison with PSMA-11 and PSMA-617 has limitations. The low patient numbers might introduce errors, especially on the highly variable inter-individual tumor uptake. In addition, due to the higher positron energy and increased spill-out-effects, 68Ga quantification might be systematically underestimated in small lesions. Furthermore, the patient selection differs: PSMA-617 was evaluated in a heterogeneous group of patients , whereas PSMA-11 was evaluated primarily in BCR patients . In contrast PSMA-1007 is now evaluated in patients with treatment-naive high-risk PCa. Therefore, quantification per SUV should be interpreted cautiously. Semi-quantitatively the ratio tumor-to-parotid is comparable for all tracers. PSMA-617 has the lowest uptake in liver, kidney and spleen, which was a goal when developing theranostic PSMA-ligands for radionuclide therapy. However, as known from 177Lu-PSMA-617 dosimetry , PSMA-617 has slower tumor accumulation than PSMA-11 and is considered suboptimal when used as 68Ga-labelled PET-tracer. 18F-PSMA-1007 shares the slightly slower tracer kinetics of PSMA-617, although this is less crucial due to the longer half-life of 18F in comparison to 68Ga. Due to its structural similarity to PSMA-617, it has been suggested that it could be an ideal diagnostic surrogate for patients considering 177Lu-PSMA-617 therapy . For a pure diagnostic tracer the slightly higher uptake of 18F-PSMA-1007 in visceral organs is negligible with regard to radiation burden and is also of limited clinical impact as visceral metastases occur late in the course of the disease. More importantly the increased uptake in tumor tissue compared to other tracers improves tumor-to-background ratios making it easier to detect small lymph node metastases. It remains to be seen whether patients with advanced castration-resistant PCa will have higher liver background which may interfere with the detection of metastases in comparison to the actual reference compound 68Ga-PSMA-11. One clear advantage for local staging is that 18F-PSMA-1007 is temporarily retained in the kidney parenchyma. For 18F-PSMA-1007, clearance via the urinary tract was only 1.2 % injected activity during 0–2 h and another 0.7 % 2–4 h p.i.; in comparison, 11 % of 18F-DCFPyl is eliminated during the first 2 h via the urinary tract and another 5 % until 3 h p.i. . Bladder content of up to SUVmax 40 (68Ga-PSMA-617) and SUVmax 100 (68Ga-PSMA-11) has been reported for the 68Ga-containing compounds, respectively [29, 30, 31, 32]. In contrast, the content of the urinary bladder was SUVmax 5 for 18F-PSMA-1007 (Fig. 4). Obviously, this agent demonstrates delayed urinary excretion and thus fulfills some of the design criteria for the ideal PSMA-targeted PET tracer. This also explains the lower radiation dose to the urinary bladder wall in comparison to other PSMA-tracers, while kidney dose is comparable (Table 2).
In the evaluation of first patients, the sensitivity of 18F-PSMA-1007 for small lymph node metastases was approximately 95 %. Metastatic nodes as small as 1 mm in diameter were discovered. In a retrospective analysis of patients who received pelvic lymphadenectomy after imaging with a variety of PSMA-targeted tracers, the sensitivity for these very small nodes was limited . Thus, small lesion detection is a very promising early result for the new compound. However, the mentioned study suffered from several limitations, including a long interval between imaging and surgery or the lack of standardization in regard to imaging protocols and documentation of findings . Therefore, it cannot be concluded that the higher sensitivity in our cohort is caused solely by the improved tracer. Nevertheless, lymph node metastases with median diameters of 5 mm are close to the technical resolution limits of PET with 68Ga-PSMA tracers and, therefore, it would be reasonable that 18F-PSMA tracers might perform at a higher level.
18F-PSMA-1007 is an attractive alternative to 68Ga-PSMA-11 and 18F-DCFPyL. Radiation burden of a 200–250 MBq injection translates to 4.4–5.5 mSv effective dose, similar to other established PET tracers. 18F-PSMA-1007 can be produced in large amounts per batch in PET radiopharmacies with an on-site cyclotron, reducing the demand for multiple tracer syntheses per day and enabling transfer to satellite centers. Preliminary data suggest reduced urinary excretion and high tumor to background ratios contribute to exceptional sensitivities including for tiny tumor deposits in the body. Larger trials with this PET tracer are expected to further define its capabilities and role in the management of prostate cancer.
We wish to acknowledge Dr. Peter Choyke from the National Cancer Institute who helped edit the manuscript for clarity, and Karl Schmidt from ABX-CRO GmbH for technical assistance with the QDOSE software. We kindly appreciate also the support of Viktoria Reiswich and Kirsten Kunze in regard to dosimetric imaging acquisition as well as Yvonne Remde for her engagement in 18F-PSMA-1007 productions.
Compliance with ethical standards
Conflict of interest
Patent application for PSMA-617 for KK and UH. Patent application for PSMA-1007 for JC, MS, UH, FLG, KK. The other authors declare that they have no conflict of interest.
All patients were informed about the in-house production of the novel radiopharmaceutical, according to the German Pharmaceuticals Act §13(2b), and gave written informed consent.
Statement of human rights
We retrospectively report observations from clinical practice. For this type of study no formal clinical trial registration is required. Our institutional ethics committee approved (permit S-321/2012) in accordance with our national regulations and to the updated version of the Helsinki declaration.
Part of the pre-clinical PSMA-1007 tracer development has been granted to JC as a Post-Doc position by ABX advanced biochemical compounds.
- 1.Afshar-Oromieh A, Avtzi E, Giesel FL, Holland-Letz T, Linhart HG, Eder M, et al. The diagnostic value of PET/CT imaging with the (68)Ga-labelled PSMA ligand HBED-CC in the diagnosis of recurrent prostate cancer. Eur J Nucl Med Mol Imaging. 2015;42(2):197–209. doi:10.1007/s00259-014-2949-6.CrossRefPubMedGoogle Scholar
- 3.Morigi JJ, Stricker PD, van Leeuwen PJ, Tang R, Ho B, Nguyen Q, et al. Prospective comparison of 18F-Fluoromethylcholine versus 68Ga-PSMA PET/CT in prostate cancer patients who have rising PSA after curative treatment and are being considered for targeted therapy. J Nucl Med. 2015;56(8):1185–90. doi:10.2967/jnumed.115.160382.CrossRefPubMedGoogle Scholar
- 4.Afshar-Oromieh A, Zechmann CM, Malcher A, Eder M, Eisenhut M, Linhart HG, et al. Comparison of PET imaging with a (68)Ga-labelled PSMA ligand and (18)F-choline-based PET/CT for the diagnosis of recurrent prostate cancer. Eur J Nucl Med Mol Imaging. 2014;41(1):11–20. doi:10.1007/s00259-013-2525-5.CrossRefPubMedGoogle Scholar
- 6.Sterzing F, Kratochwil C, Fiedler H, Katayama S, Habl G, Kopka K, et al. (68)Ga-PSMA-11 PET/CT: a new technique with high potential for the radiotherapeutic management of prostate cancer patients. Eur J Nucl Med Mol Imaging. 2016;43(1):34–41. doi:10.1007/s00259-015-3188-1.CrossRefPubMedGoogle Scholar
- 9.Giesel FL, Sterzing F, Schlemmer HP, Holland-Letz T, Mier W, Rius M, et al. Intra-individual comparison of 68Ga-PSMA11-PET/CT and multi-parametric MR for imaging of primary prostate cancer. Eur J Nucl Med Mol Imaging. 2016;43(8):1400–6. doi:10.1007/s00259-016-3346-0.CrossRefPubMedPubMedCentralGoogle Scholar
- 10.Fendler WP, Schmidt DF, Wenter V, Thierfelder KM, Zach C, Stief C, et al. 68Ga-PSMA-HBED-CC PET/CT detects location and extent of primary prostate cancer. J Nucl Med. 2016.Google Scholar
- 11.Vallabhajosula S, Nikolopoulou A, Babich JW, Osborne JR, Tagawa ST, Lipai I, et al. 99mTc-labeled small-molecule inhibitors of prostate-specific membrane antigen: pharmacokinetics and biodistribution studies in healthy subjects and patients with metastatic prostate cancer. J Nucl Med. 2014;55(11):1791–8. doi:10.2967/jnumed.114.140426.CrossRefPubMedGoogle Scholar
- 13.Cho SY, Gage KL, Mease RC, et al. Biodistribution, tumor detection, and radiation dosimetry of 18F-DCFBC, a low-molecular-weight inhibitor of prostate-specific membrane antigen, in patients with metastatic prostate cancer. J Nucl Med. 2012;53(12):1883–91. doi:10.2967/jnumed.112.104661.CrossRefPubMedPubMedCentralGoogle Scholar
- 17.Kratochwil C, Bruchertseifer F, Giesel FL, et al. 225Ac-PSMA-617 for PSMA targeting alpha-radiation therapy of patients with metastatic castration-resistant prostate cancer. J Nucl Med. 2016.Google Scholar
- 21.Cardinale J, Schäfer M, Benešová M, et al. Preclinical evaluation of 18F-PSMA-1007: a new PSMA-ligand for prostate cancer imaging. J Nucl Med. 2016. doi:10.2967/jnumed.116.181768.
- 27.ICRP: basic anatomical and physiological data for use in radiological protection reference values. ICRP Publication 89. Ann. ICRP. 2002; 32 (3–4): 1–277.Google Scholar
- 29.Afshar-Oromieh A, Malcher A, Eder M, et al. PET imaging with a [68Ga]gallium-labelled PSMA ligand for the diagnosis of prostate cancer: biodistribution in humans and first evaluation of tumour lesions. Eur J Nucl Med Mol Imaging. 2013;40(4):486–95. doi:10.1007/s00259-012-2298-2.CrossRefPubMedGoogle Scholar
- 30.Afshar-Oromieh A, Hetzheim H, Kratochwil C, et al. The theranostic PSMA ligand PSMA-617 in the diagnosis of prostate cancer by PET/CT: biodistribution in humans, radiation dosimetry, and first evaluation of tumor lesions. J Nucl Med. 2015;56(11):1697–705. doi:10.2967/jnumed.115.161299.CrossRefPubMedGoogle Scholar
- 31.Pfob CH, Ziegler S, Graner FP, Köhner M, Schachoff S, Blechert B, et al. Biodistribution and radiation dosimetry of 68Ga-PSMA HBED CC-a PSMA specific probe for PET imaging of prostate cancer. Eur J Nucl Med Mol Imaging. 2016.Google Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.