1
Dosimetry For
177Lu-PSMA Therapies
Sajeda Alawi
Master’s thesis
Master’s degree in medical physics programme University of Eastern Finland
Department of Applied Physics May 2021
2 University of Eastern Finland, Faculty of Science and Forestry
Department of Applied Physics
Master’s Degree in Medical Physics Programme Alawi, Sajeda: Dosimetry for 177Lu-PSMA Therapies Master of Science Thesis, 55 pages.
Supervisors: Ari-Petteri Ronkainen, PhD Tiina Marja Laitinen, PhD and Petro Julkunen, PhD.
May 2021.
Keywords: dosimetry, MIRD, voxel, targeted radionuclide therapy, prostate cancer, mCRPC, PSMA, Lutetium.
Abstract
Dosimetry is a necessary process in radiation therapies to ensure required energy for effective treatment is transferred to the target while saving the healthy organs. Dosimetry using MIRD formalism has been used clinically but it has limitations as it is a population-based method. Voxel-level dosimetry methods have been developed to introduce patient-specificity to dosimetry, but are still mainly under research with a few commercial ready solutions to be applied for targeted radionuclide therapy (TRT). One of the significant applications of TRT is in treating mCRPC patients with 177Lu-PSMA. Purpose: Estimating and comparing the absorbed doses of kidneys and liver for mCRPC patients post 177Lu-PSMA therapy using MIRD and voxel-level dosimetry. Methods: 12 patients went through 1 to 5 cycles of 177Lu-PSMA treatment and were imaged with SPECT-CT in Kuopio University Hospital at 4h, 24h and possibly 48h time points.
Then, their kidney and liver doses were estimated using HERMES software. Results: Absorbed doses estimated with voxel-level dosimetry were higher than absorbed doses estimated with MIRD. The mean absorbed doses for kidneys and liver based on MIRD dosimetry were 3.29 and 0.219 Gy, respectively.
Respective values based on voxel-level dosimetry were 7.91 and 0.869 Gy. The accumulative doses for all patients were within established tolerance limits for kidneys and liver. The highest kidney dose was determined for patient who had only one functional kidney. Conclusion: Absorbed doses estimated with voxel-level dosimetry were higher than the ones estimated with MIRD dosimetry mostly due to the difference in estimation of time activity curves (TACs). In voxel-level dosimetry software, a physical decay of the radioisotope was assumed after the final sampled activity point of the TAC, whereas biological decay was used in MIRD dosimetry. Thus, when using voxel-level dosimetry with the studied software, one should always image a late time point (e.g. 1 week after treatment) to estimate the true decay and extraction of 177Lu from the body.
3
Acknowledgments
I would like to express my sincere gratitude for my supervisors Ari-Petteri Ronkainen, Tiina Marja Laitinen and Petro Julkunen, for offering me this research opportunity and never hesitate to provide their endless support, in teaching me new aspects and for their practical and academic guidance throughout this journey.
I would like to thank all University of Eastern Finland’s staff and teachers for maintaining our studies despite the pandemic situation.
I am grateful to my parents, Kameela and Sayed Hasan, who believed in the power of knowledge and for their endless support by taking care of my child, Sayed Adam. I am thankful to my husband, Sayed Hasan, for his love and encouragement.
Above all, I would like to praise and thank Allah, The Most Beneficent and The Most Merciful, for surrounding me with his blessings.
4 Table of Contents
1. Introduction ... 5
2. Background Information ... 6
2.1 Nuclear Medicine ... 6
2.2 Atomic and Nuclear Physics ... 6
2.2.1 Basics of Atoms ... 6
2.2.2 Basics of Radiation ... 9
2.2.3 Radioactive Decays ... 10
2.3 Radionuclide Production ... 13
2.4 Lutetium-177 ... 15
2.5 Interaction of Radiation with Matter ... 16
2.6 Radiation Detectors ... 18
2.6.1 Gas Filled Detectors ... 19
2.6.2 Scintillation Detectors ... 19
2.7 Gamma Camera ... 19
2.8 Single Photon Emission Computed Tomography (SPECT) ... 22
2.8.1 SPECT Acquisition ... 22
2.8.2 SPECT Image Reconstruction ... 22
2.8.3 Quantitative SPECT ... 26
2.9 Positron Emission Tomography (PET) ... 29
2.10 Radiation Dosimetry ... 30
2.10.1 The MIRD Dosimetry ... 32
2.10.2 Voxel Level Dosimetry ... 33
2.11 177Lu-PSMA Therapies ... 34
2.12 Radiation Biology ... 35
3. Materials and Method ... 37
4. Results ... 41
5. Discussion ... 48
6. Conclusions ... 50
7. References ... 51
5
1. Introduction
Since their discovery by Wilhelm Roentgen in 1895, x-rays have been used widely for medical and non- medical purposes. This was followed by the identification of higher energy waves of the wave spectrum, gamma rays (International Atomic Energy Agency, 2005). Gamma rays have been used for cancer therapy due to their ability to penetrate human tissues, while transferring energy and causing damage to the cell’s DNA. Damaging DNA can cause cell reproductive failure, division delay or mutation. These effects can be controlled by adjusting the radiation energy administered to patients according to the purpose, imaging or cancer therapy (Vens et al., 2018). Radiation applied to the patient has to be as low as reasonably achievable (ALARA), to achieve the desired benefit and reduce the risk to the minimum (International Atomic Energy Agency, 2005).
The effect of ionizing radiation in damaging the DNA of human cells, can be used as a non-surgical cancer treatment or as a palliative treatment. Tumours can be exposed to ionizing radiation with multiple modalities. In external beam radiation therapy (EBRT), the tumour is exposed to radiation source that is outside of the patient body (Hoskin, 2019). Then, there are two forms of internal radiotherapy. In brachytherapy, a solid radiation source is inserted inside the patient body, inside or next to the tumour (Han et al., 2014). Another way of internal radiotherapy is targeted radionuclide therapy (TRT), or systemic radiation therapy (SRT), where radiation is delivered to the target by a radiopharmaceutical (McDougall, 2000).
As said, targeted radionuclide therapy uses unsealed radiation sources that are introduced to the patient orally or parenterally. Then, the radiopharmaceutical accumulates and stays in the target organ long enough to deliver a damaging radiation dose to the cancerous cells (McDougall, 2000). In this process, a part of the radiopharmaceutical will end up in non-targeted organs. So besides achieving the therapy goals, it is important to avoid healthy organ toxicity to assure patient safety (McDougall, 2000). This can be achieved by measuring or estimating the absorbed radiation energy in patients' organs, which is a process called dosimetry (Greene et al., 2018).
Dosimetry, in nuclear medicine procedures, is conventionally done using Medical Internal Radiation Dose (MIRD) formalism. MIRD formalism uses constant S-values to estimate the absorbed dose in a target organ from an activity in a source organ, based on anthropomorphic phantoms (Hindorf, 2014). Quantitative imaging in nuclear medicine has developed and raised the interest in more detailed voxel-based dosimetry for TRT, where the absorbed dose depends on patient-specific characteristics and metabolism (Tran-Gia et al., 2020).
One of the efficient TRT procedures is prostate-specific membrane antigen (PSMA) directed radioligand therapy for patients with metastatic castration-resistant prostate cancer (mCRPC) (Xue et al., 2020). In this study, MIRD formalism and voxel-based dosimetry are performed for patients with mCRPC, who have been treated with Lutetium-177 prostate-specific membrane antigen (177Lu-PSMA) therapy in Kuopio University Hospital. The 177Lu-PSMA treatments were started in the hospital during spring 2020 and the routine dosimetry has been performed using MIRD formalism. The aim of this work was to become familiar with
6
the voxel-based dosimetry, compare the results with both dosimetry methods and examine needs to develop the dosimetry routine.
2. Background Information 2.1 Nuclear Medicine
In nuclear medicine, a pharmaceutical compound labelled with radioactive source called radionuclide (radiopharmaceutical) is administered to patient’s body. Then the radionuclide activity is traced, as it accumulates in different concentrations in patient’s organs, using external imaging detectors (Cherry et al., 2012). External imaging detectors can be sequential planar scintillation camera, single photon emission computed tomography (SPECT) or positron emission tomography (PET) (Seco and Verhaegen, 2013). The radiopharmaceuticals can be used for both imaging and radionuclide therapy purposes (Sgouros, 2019).
The advantage of nuclear medicine technique is that its output is functional information of the organ, not only anatomical information, without interfering with organ’s physiology. Added to that, it gives molecularly the most sensitive results compared to the other imaging modalities, as it detects masses up to picomolar level. Most organ functions can be studied by labelling the radioactive source with different biological molecules. Detected images can be analysed in direct way or in quantitative way, in a relatively short time (De Lima and Webster, 2010).
On the other hand, applications in nuclear medicine are limited by poor spatial resolution, which can be handled by applying another imaging modality such as computed tomography (CT). Also, the application of radionuclides inside patients requires more indications about radiation protection because the signal detected from active source unlike in conventional x-ray, where the detected signal is accumulative from passive source (De Lima and Webster, 2010).
2.2 Atomic and Nuclear Physics 2.2.1 Basics of Atoms
The atom structure needs to be introduced to understand the radiation phenomena. First, atoms are the smallest unit in which the chemical element can be broken down without losing its chemical characteristics. Atoms can combine chemically forming compounds and bigger macroscopic structures.
Through centuries, atomic models have gone through different developments, but now it is established that the structure of an atom is a nucleus that contains positively charged protons and neutral neutrons, surrounded by negatively charged electrons that revolve around the nucleus in distinct orbits (Lilley, 2001), shown in Figure [1]. The number and distribution of electrons define the chemical characteristics of the
7
element, whereas the number of protons and neutrons are defining atom stability and electron configuration. Atoms are neutral with equal number of electrons and protons (Cherry et al., 2012).
Figure [1]: Atomic structure shows nucleus containing protons and neutrons, surrounded by schematic electron orbits (Figure adapted from Fig 1.5 in Powsner et al. 2013).
Atomic model introduced by Bohr in 1913, defines the distribution of electrons surrounding the nucleus to be quantized in diameter of shells, where the innermost shell (n = 1) is called the K shell, then the L shell (n = 2), the next is the M shell (n = 3), N shell (n = 4), and so on (Lilley, 2001). The number of electrons in each shell is limited to a maximum of 2n2 (Powsner et al., 2013).
Pauli Exclusion Principle in 1925 developed Bohr model by modifying the electrons behaviour in their orbits. It states that no two orbital electrons have the same motion as they have different spin orientation (Powsner et al., 2013).
The most stable configuration for the atom is when the innermost shells are filled with electrons. Electrons can be moved to higher shells or be kicked out of the atom when they receive surplus energy. Electron binding energy is defined as the energy needed by the electron to be removed from a certain shell and it equals the difference in binding energies between the two shells of interest (Lilley, 2001).
If electron gains energy, it can move to a higher energy state, in higher shell, but this is an unstable state, and to achieve stability, the electron tends to return to its original lower energy state. In the process of returning, the electron emits its excess energy in form of electromagnetic radiation called photon. For each element, electron binding energy differences have equal characteristic values, so the photon emissions are called characteristic radiation or characteristic x-rays (Lilley, 2001).
Auger electron might be emitted instead of a photon, when the excited electron returns to its original position. This occurs, if it transfers its energy to another electron in the outer shell, causing its ejection from its orbit. This is called the Auger effect. Elements have different probabilities to emit x-rays or Auger electrons. Atomic number of the element can affect the possibilities as heavy elements with high atomic number tend to emit x-rays and light elements with low atomic number tend to emit auger electrons (Pryma, 2014).
8
Beside orbital electrons, nucleus has nucleons, protons and neutrons, where two type of forces appear in the short range of nucleus. First force is between the positively charged protons, known as repulsive Coulombic or electrical forces. Second forces, called nuclear forces, appear between any two nucleons and help in holding the nucleus together (Powsner et al., 2013).
The motion of nucleons is limited by the described forces and their motion follows shell model, which explain the motion of nucleons in orbits that are identified by nuclear quantum numbers. Stability state of the nucleon is known as the ground state. Unstable states are excited state, when nucleons have transient existence before converting into a different state, and metastable state or isomeric state, where it has a longer lifetime before converting to a different state. To identify the states of elements, asterisk notation is used for excited states (AX*) and letter m is used for metastable states (AmX or X-Am), where (A) is the mass number of element, which represents the total number of protons and neutrons in the nucleus (Cherry et al., 2012).
Same as electrons, the nucleons can absorb excess energy, which may lead to their excitation and further to return to their ground state or stability by emitting photons of electromagnetic radiation or particles such as electrons or alpha (α) particles. This process is called radioactive decay. The emitted photons are higher in energy than the ones emitted by electron excitation and are called γ-rays. Binding energy between two different states can be calculated by Einstein's famous equation as
𝐸 = 𝑚𝑐2, [1]
which describes the conversion of mass m to energy E, and c is the speed of light that approximately equals 3 × 108 𝑚
𝑠 (Lilley, 2001).
In nature, the nuclides can be found in their stable and unstable state, where stable nuclides can be artificially transformed to their unstable state, to be radioactive nuclides. Nuclides are categorized in three groups according to the relation between their neutrons number (N) and atomic number (Z). First are isotopes, which are same elements as they have the same atomic number, but different mass number due to a difference in neutron number. Second are isobars, which are elements with the same mass number but with a different atomic number. Third are isotones, which are elements with the same number of neutrons and a different mass number (L'Annunziata, 2012). All nuclides arranged according to their number of neutrons and protons in Figure [2] showing the stable nuclides on a line called the line of stability (Powsner et al., 2013).
9 Figure [2]: Nuclides arrangement relative to their number of neutrons (N) and protons (P), (Figure
adapted from Figure 1.10 in Powsner et al. 2013).
2.2.2 Basics of Radiation
Radiation is defined as travelling energy that is emitted from a source. It can be particulate radiation, where energy is carried as kinetic energy of moving particles like electrons or protons. Second form of radiation is electromagnetic radiation, where energy is carried by electrical and magnetic fields that propagate with the speed of light. Both types can be generated from radioactive decays of atoms (Cherry et al., 2012).
The familiar type of radiation is electromagnetic radiation, where different electromagnetic radiation energies are defined in spectrum. Electromagnetic spectrum starts at low energy radio waves, microwaves and visible light, which act like waves in their interactions with matter. Followed by higher energy electromagnetic waves, such as x-rays and gamma rays, which interact with matter more as photons.
Photons are packets of energy with no mass or electrical charge and travel at the speed of light. The relation of the wavelength (𝜆) and frequency (𝑓) of electromagnetic radiation can be expressed as
𝑓 𝜆 = 𝑐, [2]
where c is the speed of light and approximately equals 3 × 108 m
s (Lilley, 2001).
Energy and mass involved in radioactivity decays are very small. Hence, units are typically modified for atomic scale and mass is presented as atomic mass units (u)
1 u = 1.66054 × 10−27 kg [3]
10 and the energy is presented as electron volts (eV)
1 eV = 1.6022 × 10−19 kg m2
sec2 . [4]
Electron volt is defined as the energy gained by a single electron when it is accelerated in electrical potential of 1 V in vacuum. The energy generated from 1 u of mass is 931.5 MeV, using equation [1]
(Powsner et al., 2013).
2.2.3 Radioactive Decays
Radioactive decay is the process of radiation emission from a nuclear reaction or decay of unstable atomic nuclei, where the unstable nucleus loses energy by emitting elementary particles such as alpha particles, beta particles, neutrons and gamma ray photons. In case the energy loss takes place in electron shells, Auger electrons and x-ray photons can be emitted. In decay processes, the rate of decay, known as rate of disintegration, is directly proportional to the mass of radionuclide. Thus, analysing radioactivity of radionuclide can be done by counting the radionuclide decays versus the time. Another method is to measure the mass of radionuclide with mass spectrometry (L'Annunziata, 2012).
Radioactive decay can happen naturally or artificially, to achieve atom stability when the number of protons approximately equal the number of neutrons. In radioactive decays, a parent nucleus (P) is transformed into a daughter nucleus (D). The parent nucleus differs from the daughter nucleus in atomic number (Z), neutron number (N), atomic mass (A) or in a combination of these. Added to that, one or more particle types can be released in one radioactive decay (Podgoršak, 2014).
Decay modes are named after the emitted particle in the decay process and gathered in six main categories: alpha (α), beta (β), gamma (γ), spontaneous fission, proton emission and neutron emission decays. Subcategories of beta (β) decay are beta plus (β+), beta minus (β-) decay and electron capture (EC). Subcategories for gamma (γ) decay are gamma emission and internal conversion (IC) (Podgorsak, 2014). The most significant decays in nuclear medicine are alpha, beta, gamma, electron capture and internal conversion.
Each radionuclide has its characteristic mode of decay, average lifetime and transition energy, defined as the total mass of the atom minus the energy emitted in form of mass and kinetic energy of emitted particle (Podgorsak, 2006).
Alpha decay occurs mainly in heavier elements, as the parent nucleus ejects two protons and two neutrons in form of a helium nucleus 24𝐻𝑒, called an alpha particle. Alpha particles can travel only up to 0.001 cm in soft tissues with kinetic energy of 4-9 MeV. The decay reaction can be expressed as (Podgorsak, 2006)
𝑃→𝛼
𝑍𝐴 𝑍−2𝐴−4𝐷 + 𝐻𝑒24 . [5]
11
In beta decay the parent atomic number changes by ±1 and the atomic mass does not change, hence, the daughter atom is an isobar of the parent atom. The decay can happen in three ways. First, in 𝛽−decay, where a neutron is converted into a proton, electron with kinetic energy and antineutrino expressed as (Podgorsak, 2006)
𝑍𝑃
𝐴 𝛽−
→ 𝑍+1𝐴𝐷 + 𝑒−+ 𝑣𝑒. [6]
𝛽− particles penetrate relatively small thicknesses, but thicker than alpha particles. If the produced daughter is not stable yet, it may decay and emit massless gamma ray (𝛽−, 𝛾) (Podgorsak, 2006)
𝑍𝑃
𝐴 𝛽−
→ 𝑍+1𝐴𝐷∗→𝛾 𝑍+1𝐴𝐷. [7]
Second, in positron decay 𝛽+, a proton is transformed into a neutron, positron with kinetic energy and antineutrino. Positron goes through annihilation reaction, where two gamma rays are produced with energy of 0.511 MeV for each and separated by 180° angle (Podgorsak, 2006). The process notation is (𝛽+, 𝛾) and can be expressed as
𝑍𝑃
𝐴 𝛽+
→ 𝑍−1𝐴𝐷 + 𝑒++ 𝑣𝑒. [8]
The third way is EC, which has similar effect on the parent atom as 𝛽+decay 𝑃 + 𝑒−
𝐴𝑍 𝐸𝐶
→ 𝑍−1𝐴𝐷 + 𝑣𝑒. [9]
As shown in the formula, in EC the atom captures an electron and releases a neutrino with kinetic energy.
𝛽+and EC are competing processes, with 𝛽+decay being more probable in light elements and EC in heavier elements; because electrons are closer to the nucleus and more readily captured (Podgorsak, 2006).
In gamma decay excessive energy is emitted in form of high energy massless gamma rays. Another form of gamma decay is internal conversion, where the produced daughter in alpha and beta decays might not be stable. Metastable states tend to become stable by going through isomeric transition by emitting gamma rays or by transferring the extra energy to an orbital electron expressed as
𝐷∗
𝐴𝑍 → 𝐷𝑍𝐴 ++ 𝑒−+ 𝑄𝐼𝐶 → 𝐷,𝑍𝐴 [10]
where atom D in metastable state 𝐴𝑍𝐷∗ goes through internal conversion, where 𝑍𝐴𝐷+ ionized state of atom D resulted with emitted electron and decay energy 𝑄𝐼𝐶. Followed by de-ionization of atom D to reach its stability (Podgorsak, 2006).
As there are multiple competing decay processes with different probabilities the decays cannot be predicted for individual atoms. Therefore, radioactive decay parameters are averaged for a bunch of atoms. Decay rate is the number of disintegrations or decays per time unit
𝐴𝑐 = −𝑑𝑁𝑜
𝑑𝑡 , [11]
where Ac is the activity or decay rate in Becquerel (Bq), dNo is the change in the number of atoms, dt is the time in seconds. The minus sign refers to the decrease in the number of atoms in the decay process.
12
The activity unit is Becquerel which is the number of disintegrations per second and relates to an older used unit curie (Ci) as
1 mCi = 37 MBq. [12]
The radioactivity given to patients in nuclear medicine procedures is given usually in MBq or mCi and is called dosage that is to be differentiated from the dose term (Chandra, 2004).
The instantaneous activity has been found to be dependent on the number of atoms No present at time t and on a constant value identified for each radionuclide separately called characteristic decay constant (λ)
𝐴𝑐 = 𝜆 𝑁𝑜. [13]
The unit of λ is 1/s. If radionuclide decays in different modes, the decay constant λ is different in each mode and can be determined as their sum (Chandra, 2004).
λ = λ1+ λ2+ λ3+ ⋯ [14]
The time it takes for the number of atoms to decay to their half amount is called half-life (𝑇1
2
). Time activity curve (TAC) can be plotted between activity Ac(t) versus time t, which shows an exponential decay of activity
𝐴𝑐(𝑡) = 𝐴𝑐(0)𝑒
−ln 2 𝑡 𝑇1
2 , [15]
where Ac(t) is activity at time t, Ac(0) is activity at t=0. Decay equation can be written with respect to the decay constant λ as
𝐴𝑐(𝑡) = 𝐴𝑐(0)𝑒−λt, [16]
where decay constant λ =ln 2
𝑇1 2
(Powsner R. and Powsner E., 2006).
The exponential term is called decay factor (DF)
𝐷𝐹 = 𝑒−λt. [17]
The average lifetime of radionuclide (𝜏) is the reciprocal of decay constant 𝜏 = 1
λ=
𝑇1 2
ln 2. [18]
(Powsner R. and Powsner E., 2006).
The described activity formulas are true for the physical decay outside the human body. Inside the human body, the radionuclide is also excreted biologically, which is characterized by the biological half-life of radionuclide 𝑇𝑏. The biological half-life is the time it takes for half of the radionuclide to exit the biological system. As both the biological and the physical decays are concurrently happening, their combined effect
13
must be considered when choosing suitable time for patient imaging after the pharmaceutical administration. The effective half-life 𝑇𝑒𝑓𝑓 is given as,
1 𝑇𝑒𝑓𝑓 = 1
𝑇𝑝+ 1
𝑇𝑏, [19]
where Tp is the physical half-life and Tb is the biological half-life (Chandra, 2004).
2.3 Radionuclide Production
Radionuclides or radioisotopes are atoms in unstable state and tend to emit their excess energy to achieve stability. Each radioisotope is characterized by its half-life, mode of its decay, the energy of produced emissions, its specific activity, purity, chemical properties and the cost. Beside the characterized properties of the radionuclide, its biological targeting and clearance from patient body should be considered, especially when labelling pharmaceuticals with radioisotopes that may affect its chemical property and organ physiology (Reichardt and Bacchi, 2005).
Short half-life radionuclides are produced for medical purposes, for imaging or therapy, by bombarding the nucleus of stable atom with sub-nuclear particles. This production process can be performed with nuclear reactors or particle accelerators. Radionuclides can also be extracted from radionuclide generators that utilize longer half-life parent nucleus (Cherry et al., 2012).
Nuclear reactors can produce large quantities of radionuclides. Most commonly, the core of nuclear reactor contains natural uranium (235𝑈) with half-life of approximately 7 × 108 years, where nuclear fission is initialized by bombarding the uranium-235 with neutrons, which causes the uranium to split into two lighter elements and 2-5 neutrons. Resulted neutrons can cause nuclear fission reaction again, forming a nuclear chain reaction, such as
𝑈 + 𝑛 → 23692𝑈∗ → 14456𝐵𝑎
23592 + 𝐾𝑟 + 3𝑛3289 , [20]
(Cherry et al., 2012).
The produced elements from nuclear fission usually have extra neutrons and tend to emit energy through 𝛽− decay until they reach stability, for example in Molybdenum-99 production
𝑌𝛽
− (1.5 𝑠)
→
3999 𝑍𝑟4099 𝛽
− (21 𝑠)
→ 𝑁𝑏 𝛽
− (15 𝑠)
→
4199 4299𝑀𝑜. [21]
Produced radionuclides with suitable half-life can be extracted from the process chain. Radionuclides from fission fragments are carrier-free and have high specific activity (Cherry et al., 2012).
Molybdenum is used widely in radionuclides generators, which are used to produce radionuclides by having long half-life parent radionuclide that decays into a shorter half-life daughter radionuclide that is extracted when needed. In Technetium generators, the daughter nuclide technetium 99𝑚𝑇𝑐 with half-life of 6 hours is extracted from the parent nuclide molybdenum 99𝑀𝑜 that has half-life of 66 hours. In this
14
technique, partial elution of the long half-life parent radionuclide could occur in the extraction, resulting in a higher dose delivered to the patient, but this usually very minimal (Bushberg et al., 2011).
Another method to produce radionuclides is to accelerate charged particles, such as protons, deuterons (12𝐻) and alpha particles, to very high energy (10-20 MeV) and bombard the element’s nucleus to an excited state. The two common reactions of this type are the proton capture
𝑋 (𝑝, 𝑛) 𝑍+1𝐴𝑌
𝑍𝐴 [22]
and the deuteron capture reactions
𝑋 (𝑑, 𝑛) 𝑍+1𝐴+1𝑌
𝐴𝑍 . [23]
Different types of accelerators have been developed for this purpose such as Van de Graaf, linear accelerators and cyclotrons that are the most common accelerators used to produce radionuclides for nuclear medicine. The radionuclides produced by cyclotrons are of smaller quantities, carrier free, short half-life and decay mostly by EC or 𝛽+decay (Cherry et al., 2012).
Cyclotron is an accelerator composed of two hollow D-shape electrodes (Figure [3]), which produce an electric potential to accelerate the charged particles produced from the centre of the cyclotron. Constant magnetic field is present to guide the circular path of the accelerated charged particles. Positively or negatively charged particles can be generated and accelerated in circular path and finally extracted to bombard the target to generate the radionuclide of interest (Strijckmans, 2001).
Figure [3]: Main structure of cyclotron (Figure adapted from Figure 16.1 in Bushberg et al., 2011).
Cyclotron can generate a wide variety of radionuclides such as Carbon-11, Nitrogen-13, Oxygen-15 and Fluorine-18 (Bushberg et al., 2011). Fluorine-18 has a major role in diagnosis as it can detect abnormal glucose metabolism in all organs if used in form of fluorodeoxyglucose called FDG, which is 2-[18F]fluoro- 2-deoxy-D-glucose (International Atomic Energy Agency, 2012).
FDG production starts in cyclotron where Oxygen-18 is bombarded with protons to produce radioactive Fluorine-18, which is subsequently transferred to a radiopharmaceutical laboratory, where it is further prepared to be suitable for clinical use in the form of FDG (International Atomic Energy Agency, 2012).
15
Neutron activation is another method used to produce radionuclides in nuclear reactors by using the produced neutrons to excite the nucleus of chosen elements. Capturing neutrons can cause emission of gamma rays
𝑍𝑋
𝐴 (𝑛, 𝛾)𝐴+1𝑍𝑋, [24]
or cause emission of proton
𝑍𝑋
𝐴 (𝑛, 𝑝)𝑍−1𝐴𝑌. [25]
Unstable produced elements from neutron activation are small in amount, non-carrier free and decay by 𝛽+to reach stability (Cherry et al., 2012).
Radiopharmaceutical preparation aims to produce pure or called carrier-free radionuclides. The presence of carriers, which are unstable nuclides, but have long half-life can transfer to patient body and expose it with redundant radiation. The carrier free specific activity (CFSA) is calculated for the administered radiopharmaceutical as
𝐶𝐹𝑆𝐴 (Bq
g) ≈4.8×1018
𝐴×𝑇1 2
[26]
where A is atom mass number in grams and 𝑇1
2
is half-life in days. Added to that, the number of carriers present in the sample affects the time it will take for the radionuclide to distribute in patient tissues.
Furthermore, the labelling process of radiopharmaceutical is affected by the specific activity (Chandra, 2004).
2.4 Lutetium-177
Lutetium-177 notated as 17771𝐿𝑢 is an unstable radioisotope. It is produced artificially in nuclear reactors since it cannot be found in nature due to its short half-life of 6.71 days (L'Annunziata, 2012). The production process occurs by bombarding 17671𝐿𝑢 or Ytterbium-176 176𝑌𝑏 with neutrons to produce 17771𝐿𝑢 and gamma rays notated as
𝐿𝑢 (𝑛, 𝛾)177𝐿𝑢
176 [27]
or
𝑌𝑏 (𝑛, 𝛾)177𝑌𝑏 𝛽−
176 177𝐿𝑢 [28]
respectively (Dash et al., 2015).
The Lu-177 precursor used in this study was produced in nuclear reactor by the second shown process by bombarding Yb-176 target with neutrons.
16
The unstable Lutetium-177 goes through 𝛽− decay to achieve stability and disintegrates to three excitation levels and ground state of hafnium-177 (17772𝐻𝑓), which produce different types of radiation (L'Annunziata, 2012), shown in Figure [4].
Figure [4]: Lu-177 decay diagram (Figure adapted from Figure 1 in Dash et al., 2015).
Lu-177 goes through direct ground state transition which leads to 𝛽− emission with different percentage of gamma energies, where 78% of Lu-177 decays with production of 497 keV energy, 9.1% produce 384 keV, 12.2% produce 176 keV, 6.6% are low energy gamma rays of 113 keV and 11% produce 208 keV energy, which are useful for dosimetry imaging. Besides that, electrons with 15 keV energy and Hf K-shell x-rays with energy of 58 keV are emitted. The gamma energies are compatible to be detected by a gamma camera in nuclear medicine procedures. Produced 𝛽−particles are effective in therapy process due to their mean range of 0.7 mm in soft tissue, which prevents irradiation of the surrounding tissues with high energies (Hoedl and Updegraff, 2015; Dash et al., 2015).
2.5 Interaction of Radiation with Matter
Radiation interacts in different ways with the matter it passes through, it can lose its energy partially or completely while causing changes in the atomic structure of the material. The interactions of charged particles (alpha and beta particles) and photons (gamma rays and x-rays) with matter are described separately (Powsner R. and Powsner E., 2006).
Charged particles interact with matter in terms of collisions due to the presence of electrical forces of attraction or repulsion between the charged particles and matter atoms (Cherry et al., 2012). For same reason charged particles lose their energy in a short distance compared to photons. Excitations occur when charged particle interacts with matter and transfers its kinetic energy to an inner shell electron causing it to be excited to an outer shell. The excited electron tends to return to its original place and loses its added energy in form of x-rays. Ionization occurs when the kinetic energy transferred from the charged particle
17
to the atom electron is high enough to eject the electron out of the atom (Powsner R. and Powsner E., 2006).
Bremsstrahlung interactions occur when the charged particles are deflected as a result of the nuclear forces and emit x-ray radiation (Powsner R. and Powsner E., 2006).
Probability of photon interaction with matter depends on the energy of the photon and on the atomic number (Z) of the matter, Figure [5]. Human tissue has low mean atomic number (Z=7.5), which makes Compton scattering interactions more probable than photoelectric interactions at low energies (less than 50 keV). In high atomic number, such as lead (Z=82), photoelectric effect interactions are dominant up to photon energies around 1 MeV. The third interaction type is pair production, which requires photon energy greater than 1.022 MeV (Powsner R. and Powsner E., 2006).
Figure [5]: Probability of photon interactions with matter dependency on incident photon energy and matter’s atomic number (Figure adapted from Figure 7.19 in Choppin et al., 2013).
Figure [6]: Most dominant photon interactions with matter (Figure adapted from Figure 7.16 in Choppin et al., 2013).
18
Photoelectric effect occurs when a photon ejects an inner shell electron, see Figure [6]. The ejected electron is called photoelectron and the vacancy of the photoelectron is filled by an outer shell electron.
The excess energy between the outer and inner shells is emitted in form of characteristic x-ray (Powsner R. and Powsner E., 2006). The energy of the ejected photoelectron can be calculated as
𝐸𝑝ℎ𝑜𝑡𝑜𝑒𝑙𝑒𝑐𝑡𝑟𝑜𝑛 = 𝐸𝑝ℎ𝑜𝑡𝑜𝑛− 𝐸𝑏𝑖𝑛𝑑𝑖𝑛𝑔. [29]
In Compton scattering interaction, the incident photon hits an atom and causes ejection of an outer shell electron, called Compton electron, see Figure [6]. Additionally, the incident photon loses its energy and changes its direction in an angle of 0° to 180° (Powsner R. and Powsner E., 2006). The scattered photon energy (𝐸𝑠𝑐) is related to its scattering angle (𝜃) and initial photon energy (𝐸𝑖) as
𝐸
𝑠𝑐=
𝐸𝑖1+( 𝐸𝑖
0.511 MeV)(1−cos 𝜃) . [30]
In pair production interaction the incident photon with high enough energy interacts with the nucleus and is converted into an electron, a positron and their kinetic energy, shown in Figure [6]. When the produced positron loses its kinetic energy, it annihilates with an atomic electron into two opposite directed gamma rays with energy of 0.511 MeV each (Cherry et al., 2012).
Photon interactions with matter decrease the intensity of photon fluence, which is called attenuation and the matter acts as an attenuator. Attenuation can be calculated as the ratio of the exit photon intensity (𝐼𝑜𝑢𝑡) to the initial photon intensity (𝐼𝑖𝑛), and it follows the same exponential form as radioactive decay, but as an exponential function of thickness of the matter (Powsner R. and Powsner E., 2006)
𝐼𝑜𝑢𝑡
𝐼𝑖𝑛
= 𝑒
−𝜇𝑥.
[31]The linear attenuation coefficient (𝜇) is a property of the matter of interaction. If the thickness (x) is expressed in cm, the unit of 𝜇 is 1/cm. The linear attenuation coefficient depends on the energy of the incident photons and on the atomic number of matter (Z). Low photon energy and large atomic number (Z) leads to greater linear attenuation coefficient values (Powsner R. and Powsner E., 2006).
2.6 Radiation Detectors
Radiation detectors are used to detect radiation activity for safety purposes, such as measuring the radioactivity of radiopharmaceuticals before administration to patients. Radiation emitted from radionuclides administered to patients needs to be detected and quantified to form a medical image. The medical images displayed give information about the energy distribution of the radionuclide, added to that time activity curve (TAC) of the radionuclide is estimated (Pryma, 2014).
Different detectors have been developed and can be categorized based on their detecting principle such as gas-filled detectors, which detect the radiation by measuring ionization of gas inside the detector
19
chamber in form of an electrical current. Another detector type is scintillation detector, which detects the radiation energy by forming visible light in the scintillation material when a photon is absorbed. The emitted light is then detected and amplified into electrical current (Cherry et al., 2012).
2.6.1 Gas Filled Detectors
In gas filled detectors, the radiation energy causes ionization of the gas inside the detector, forming ion- electron pairs that are separated with an applied electrical potential. The electrical potential causes the negative electrons to accumulate near the positive anode, whereas the ions drift to the negative cathode forming an electrical current to be read (Pryma, 2014).
Different detectors based on this ionization principle operating at different voltages have been developed.
For high dose rate radiation, ionization chamber and dose calibrator that operate at low voltage are used.
For lower dose rates, proportional counters are used. Finally, for the most sensitive detection of very low amounts of radiation, such as in case of contamination, Geiger-Mueller (GM) survey meter that operates at high voltage is used (Pryma, 2014).
2.6.2 Scintillation Detectors
Scintillation detectors contain scintillation material that can be inorganic in form of solid crystals or organic in form of liquid solutions. The scintillation material converts the energy of detected radiation into visible light. The amount of produced light is proportional to the energy of detected radiation. The small amount of produced visible light is converted into electrons bursts, which are multiplied in high voltage using photomultiplier tube (PMT) and converted to electrical current (L'Annunziata, 2012).
2.7 Gamma Camera
Gamma camera is the most basic detector used in nuclear medicine procedures to produce the medical image. Gamma camera is inorganic scintillation detector, and its major photon detector parts are collimator, scintillation crystal and photomultiplier array (Pryma, 2014), shown in Figure [7].
20
Figure [7]: Main structures of gamma camera (Figure adapted from Figure 2.1 in Brandon et al., 2014).
Collimators are composed of numerous aligned holes usually made of lead and the first part to meet the emitted photons from the patient. Collimator’s main function is to allow only photons with certain direction to enter the detector system and block the rest from detection by absorbing them in the septa between the holes, see Figure [8]. Collimators allow the image to be formed as photons are emitted from the patient in all directions (Powsner R. and Powsner E., 2006).
Figure [8]: Image of point source with collimator (on the left) and without collimator (on the right) (Figure reproduced with permission from Strömvall, 2020, p.12).
Different types of collimators can be used, such as parallel hole collimators and nonparallel hole collimators, depending on the desired direction of detection or, whether there is need to magnify or minify the organ of interest. Parallel hole collimators can be low energy all-purpose collimators (LEAP), which have relatively large holes to detect more photons. This increases the sensitivity of the collimator but lowers the resolution of the image. Another type of parallel hole collimator is high-resolution collimator, which has smaller diameter and longer length holes compared to the low energy all-purpose collimators.
This structure allows to filter the undesired directed photons from detection, which leads to decreased sensitivity and increased resolution of the image (Powsner R. and Powsner E., 2006).
Nonparallel hole collimators have angled holes such as converging collimators, which magnify the region of interest. Diverging collimators, on the other hand, can be used to minify the region of interest if it is larger than the field of view of the detector. Pinhole collimators have only one hole to allow magnification of superficial small organs of interest. Last type is fan-beam collimator which is a combination of the
21
converging and parallel-hole collimator, which is designed for rectangular gamma camera head to image small organs and allow the detected photons to cover all crystal area (Powsner R. and Powsner E., 2006).
After the collimator, the detected photons hit the crystalline scintillation material. The crystal is the scintillator, which absorb the detected photons energy and emits visible light. The common crystal used in gamma camera is sodium-iodine dipped with thallium, NaI(TI) crystal, which has a high density (3.67 g/cm3), high atomic number (Z=53) and high stopping power to absorb the photons. Another factor that affects the sensitivity of the crystal is its thickness, the thicker the crystal the more sensitive, but thicker crystal results in more weight and expenses. Besides that, thicker crystal has lower resolution as the produced light will travel in the crystal before it is detected by the photomultiplier tubes (Pryma, 2014).
The emitted visible light from the crystal is low in energy and needs to be amplified in order to produce an image, hence it enters an array of photomultiplier tubes (PMT) that amplify the signal. Photomultiplier tubes are electronic devices that convert the visible light into electrical current. They are composed of a vacuum tube with photocathode, where the visible light is absorbed and burst of electrons are produced due to photoelectric effect. Electrons are directed through a focusing electrode to series of dynodes rising in potential towards the anode. Number of electrons is amplified at each dynode and results in a detectable electrical signal at the anode that is proportional to the energy of the initially detected photons (Pryma, 2014), see Figure [9].
Figure [9]: Photomultiplier tube scheme (Figure adapted from Figure 4 in Burgess, 2005).
Finally, the energy and position of the photons are quantified through different electronic components such as, preamplifiers, amplifiers, pulse-height analysers, multichannel analyser, A/D converters and time- to-amplitude converters (Cherry et al., 2012).
22
2.8 Single Photon Emission Computed Tomography (SPECT)
Single photon emission computed tomography (SPECT) is a process of gamma camera imaging, where multiple planar images are captured from different angles to reconstruct 3D distribution of radioactive traces in the patient’s body. SPECT can be combined with radiographic and computed tomography (CT), which gives a direct correlation between the anatomical structures and the functional information of the radiopharmaceutical’s distribution (Kim and Zukotynsk, 2017).
CT is an imaging modality composed of x-ray tube with detectors on the opposite side, that rotates around the patient to acquire projections to form high resolution 3D images. The detected signals are the attenuated photon profiles around the patient, with resolution less than 1 mm and reconstructed usually into 512x512 pixel matrices (Geleijn, 2014). So, the produced image depends on attenuation coefficients of tissues, which are usually calculated to be related to a standard scale called Hounsfield units (HU) using formula
𝐻𝑈 = 1000𝜇−𝜇𝑤
𝜇𝑤 , [32]
where 𝜇 is the attenuation coefficient of imaged voxel and 𝜇𝑤 is the average attenuation of water (Geleijn, 2014).
2.8.1 SPECT Acquisition
In SPECT device, one or two gamma camera detectors with large field of view (30 x 50 cm) rotate around the patient, to allow different angle projections to be acquired. Projections can be set to be taken over an arc of 180o or 360o, resulting in a series of 2D images. From this series, a 3D data is then calculated through a process of image reconstruction. The radial motion of the detectors allows it to reduce the distance between the patient and the detector to few centimetres, which improves the spatial resolution of the image. After the energy detection is done, CT imaging can be performed to acquire anatomical 3D images.
Finally, both images are fused in the software resulting reconstructed image with functional and anatomical details, shows an accurate information for diagnosis and therapy procedures (Kim and Zukotynsk, 2017).
2.8.2 SPECT Image Reconstruction
The acquired images from different projections around the patient are used to determine the activity distribution inside the patient. Different methods are used to reconstruct images for this purpose, such as analytical and iterative methods (Strömvall, 2020).
23
Analytical methods rely on filtered back projections (FBP) techniques, where each acquired projection is filtered to remove blurring effect and added back to the acquired angle to generate trans-axial images.
The produced trans-axial images can form a 3D image, which can be viewed in any direction (Strömvall, 2020), illustrated in Figure [10].
Figure [10]: Filtered back projection method (Figure adapted from Figure 5 in Nygren, 1997).
Iterative methods are widely used, because they generate less noise in images compared to FBP and image corrections can be applied in image reconstruction step, like attenuation, scatter, collimator and partial volume effect corrections (Strömvall, 2020).
Iterative methods depend on the probability matrix of the detected photons, generated from a voxel in patient body and a pixel in the projection image. First iteration estimates a homogeneous activity distribution within the patient or a blank image. Followed by iterations that compare the calculated projections with the acquired ones, which improve the first iteration. The estimated and measured projections should be matched accurately, to reduce image noise and increase the accuracy of image quantification. Filters to reduce the image noise can be used as well after all iterations (Hutton, 2011).
Another iterative method that can be used is a statistical iterative method, called maximum likelihood expectation maximization (MLEM), which maximizes a likelihood function iteratively and considers the Poisson noise in the acquired projections. To calculate a new iteration update, it includes all the previous acquired projections. It is a stable method but slow. To make it faster, sets of acquired projections are grouped into smaller subsets and separately reconstructed. This is called ordered-subsets expectation maximization (OSEM), which is the commonly used method nowadays (Hutton, 2011).
In OSEM, subsets are grouped in evenly spread projections surrounding the patient and two projections represents the absolute minimum number per subset. Usually four projections per subset is used, where the sum of counts in all subsets should be equal and it is easily can be applied in larger subset sizes (Strömvall, 2020). Comparison between three image reconstruction methods is shown in Figure [11], where OSEM method provides images with lower noise, higher contrast and resolution compared to images reconstructed by FBP, and faster compared to MLEM method.
24 Figure [11]: Reconstructed images for phantom with 1, 2, 4 and 8 million counts using FBP (row B),
MLEM (row C) and OSEM (row D) in PET/CT (Figure reproduced with permission from Kastis et al., Copyright 2010 to Institute of Electrical and Electronics Engineers).
Further corrections can be implemented after image reconstruction using iterative methods like, attenuation correction, which corrects the loss of counts in images as a result of attenuation in patient or any other material present in the field of view. Incident photons can lose their energy as a result of photoelectric absorption, Compton scattering lowering the detected photon energy outside the desired energy window. These detected photons with low energy have no useful information about the patient and can be eliminated (Strömvall, 2020). Phantom images reconstructed with and without attenuation correction show how attenuation correction makes the contrast more homogenous, and scatter correction improves target-background contrast (Figure [12]).
Figure [12]: SPECT reconstructed image for phantom (a) without corrections (b) with attenuation correction (c) with attenuation and scatter corrections (Kuopio University Hospital Archive).
CT images are generated by detecting the attenuated photons that pass through the patient, so they can be used in attenuation correction by calibrating the measured Hounsfield units to the attenuation coefficient of the used radionuclide and generate attenuation map (Figure [13]). Calibration curves depend on the energy of incident photons, which make it radionuclide specific method (Patton and Turkington, 2008).
25 Figure [13]: Calibration curve used to convert the Hounsfield units to attenuation coefficient (Figure
adapted from Figure 8 in Patton and Turkington, 2008).
To enable CT images to perform well in reducing attenuation effects, they have to be fused accurately from all coordinates with the SPECT images. To achieve this, patient position should be kept stable through the whole procedure. In case of minor movements while the procedure, images can be matched manually or automatically in the reconstruction software. Any errors in fusing can cause errors in the quantitative analysis, which can lead to false diagnosis (Patton and Turkington, 2008).
The effect of Compton scattering usually is not detected due to the lower energy of the scattered photons, but they can leak into the detector energy window as a result of poor energy resolution. The amount of the scattered photons detected depends on the size, density, composition of the imaged body, distribution of the administered radiopharmaceutical inside the body, incident photon energy, the energy resolution of the detector and the width of the energy window. The detected scattered photons need to be removed to enhance image quality and quantification accuracy, see Figure [14] (Strömvall, 2020).
Figure [14]: Heart perfusion image without scatter correction (left) has worse contrast compared to image with scatter correction (right) (Kuopio University Hospital Archive).
26
Scatter correction methods are developed to overcome this inaccuracy, like corrections based on subtraction of scatter images by detecting the photons in separate energy windows, but it increases image noise. Another way is to add the correction into the iterative reconstruction, where the scattering can be modelled using the detected patient images in 3D. Scatter modelling is reverse process where the detected scattered photon is tracked to its original position using line source estimation or Monte Carlo estimations (Strömvall, 2020).
More corrections can be implemented to improve the poor spatial resolution, especially when imaging small objects in SPECT (Strömvall, 2020), as shown in Figure [15]. Resolution recovery or compensation for collimator-detector response (CDR) is applied to take into account the collimator geometry, septal penetration and collimator scatter response. Two methods are commonly used to model the depth- dependent CDR. First is Gaussian and exponential function fitting to measure point source responses at different source-detector distances. Second is Monte-Carlo simulations to create depth-dependent point spread functions for all required distances (Chun at al., 2012).
Figure [15]: Left: The effect object-detector distance on image spatial resolution. Right: Assumption of imaging point source as line response (in red), but in reality, it is cone like detection, where it reduces the spatial resolution and need to be corrected by resolution recovery methods (Figure adapted from
Piccinelli and Garcia, 2016).
Additional correction utilized is the partial volume effect correction, where the adjacent regions with high and low activities are saved from being underestimated or overestimated, respectively. Different methods are used like the ones that depend on defining the organ contours in CT, and others that depend on quantification of recovery coefficient, which is the ratio between the measured activity in the image and the actual administered activity in the object (Strömvall, 2020).
2.8.3 Quantitative SPECT
The reconstructed and corrected images are either used for visual detection, such as in perfusion studies, or for quantitative tasks, where some numerical values need to be measured such as the activity
27
concentration, standardized uptake values (SUVs) or organ absorbed doses, which are calculated from time activity curves (TACs) (International Atomic Energy Agency, 2014).
These measures are calculated by providing some parameters to the software system such as sensitivity factor, which is needed to transform image counts into activity concentrations, and it is measured in sensitivity calibrations done periodically for the detectors. Other parameters are the activity administered to patient (syringe activity pre and post injection), time of radiopharmaceutical administration, patient weight and height (Dickson, 2020).
Sensitivity factor, called camera calibration factor, is determined by imaging a known source of activity that has been measured with an accurate dose calibrator. To reduce the effect of differences in attenuation and scatter corrections, a patient equivalent phantom can be imaged such as a water tank.
Image acquisition and processing should be performed similarly as in patient studies (Dewaraja et al., 2013). The calibration factor (CF) is
𝐶𝐹 = 𝐶
𝐴𝑐 𝑡, [34]
where C is counts per second, Ac is activity in MBq and t is acquisition time in seconds (Zhao et al., 2018).
The volume used for calibration study should be equivalent to the quantified patient volumes, otherwise partial volume effect corrections need to be applied (Dewaraja et al., 2013).
In this study, camera calibration was performed using a uniform cylindrical phantom, with diameter 21.6 cm and height 18.6 cm. The phantom was injected with 365 MBq of Lu-177 and data was acquired from 32 views per detector (30 s per frame) using Siemens Symbia T16. Calibration coefficient was determined to be 10.4 cps/MBq, for an OSEM reconstruction of 15 iterations and 16 subsets by a dedicated Hermes HybridRecon application, which uses attenuation, collimator and scatter corrections.
In imaging-based dosimetry, TACs are generated to quantify the changes in activity distribution in a region of interest over time. TACs shows an exponential decay of the radiopharmaceutical inside the body is integrated, then the area under the TACs integrated numerically and it is used to calculate the absorbed dose in region of interest, shown in Figure [16] (Dewaraja et al., 2012).
Figure [16]: Time activity curve for four time points and the cumulated activity is the area under the curve (Figure adapted from Figure 3 in Gape et al., 2020).
28
To determine the time integrated activities in voxel level, where more image noise and propagation noise is present, a registration between SPECT and CT images is required. Registration between SPECT and SPECT is not recommended due to the poor resolution, absence of anatomic information and presence of noise (Dewaraja et al., 2012).
In voxel-level dosimetry, dose volume histograms (DVHs) from a volume of interest are generated, which they show the percentage of voxel receiving a given amount of absorbed dose, as shown in Figure [17]
(Ljungberg and Gleisner, 2016).
Figure [17]: Dose volume histograms integrated from liver images reconstructed using OSEM method (Figure adapted from Figure 8 in Ljungberg and Gleisner, 2016).
Different platforms have been developed to implement quantitative SPECT and in this study a Hybrid Recon package from Hermes Medical Solutions was used, which requires sensitivity factor for the measured radionuclide, collimator type to quantify images, the activity injected and patient habitus data.
It offers the application of CT attenuation correction, Monte Carlo scatter correction and motion correction (Dickson, 2020). Steps of SPECT image acquisition, reconstruction and quantification are summarized in Figure [18].
29
Figure [18]: Dosimetry chain in SPECT/CT (Figure adapted from Figure 2 in Ljungberg and Gleisner, 2015).
2.9 Positron Emission Tomography (PET)
Positron emission tomography is an imaging modality, where positrons emitted from radionuclides are detected. PET detection principle is based on annihilation coincidence detection (ACD), where emitted positron has very short half-life and goes into annihilation reaction with an electron. Annihilation produces two high energy photons (E=511 keV) that are ejected in approximately opposite directions (𝜃 = 180°).
As a result of the opposite direction of emitted photons, no collimators are needed. The detection crystals Images Acquired in CT and SPECT
SPECT Images Reconstructed by Applying Corrections for Attenuation,
Scattering and Resolution Recovery
SPECT Images Quantification by Applying Calibration Factor to Convert
Measured Counts/Second to Activity
Acquisions at Multiple time Points (4, 24 and 48h posttherapy)
Image Registration between All Acquired Images
Delineation of Regions of Interest
Time Activity Curve Generated and Fitted
Determination of Mass
Calculation of Absorbed Dose
30
are arranged in a ring-shaped array surrounding the patient. In PET, applying another imaging modality is still required to enhance visualization of the anatomy, such as CT or magnetic resonance imaging (MRI) (Treglia and Giovanella, 2020), see Figure [19].
Figure [19]: Scheme of PET/CT imaging modality (Figure adapted from Figure 1 in Arora and Bhagat, 2016).
The main characteristics of PET scan compared to SPECT are higher sensitivity due to absence of collimators, better image resolution, less noise equivalent count rate, since higher energy photons produce less attenuation in the tissue, higher image contrast and relatively less time needed for imaging (Treglia and Giovanella, 2020).
PET scans require generation of positron emitting radionuclides, which are most often produced with cyclotrons. Commonly used positron emitting radionuclides are, Carbon-11 (11C) with a half-life of 20 minutes, Fluorine-18 (18F) with a half-life 110 minutes, Gallium-68 (68Ga) with a half-life of 67 minutes and Copper-64 (64Cu) with a half-life of 12.7 hours (Treglia and Giovanella, 2020). Out of these, Fluorine-18 labelled glucose called Fluorodeoxyglucose (FDG) is the most used PET tracer.
2.10 Radiation Dosimetry
Dosimetry is the process of approximating the absorbed dose, which is the energy of the ionizing radiation absorbed by the mass of tissue (Stabin, 2008). Dosimetry can be used to verify that treatment doses will be under normal tissue tolerances, hence minimizing radiation toxicity. On the other hand, dosimetry can be used to achieve planned radiation dose and biological effect in the treated tumours. It is also an important in research, like in the case of introducing new radiopharmaceuticals. Dosimetry also provides verification of the planned dose for recording purposes, for radiation protection officials or in case it is needed to make further medical decisions for the patient, like in a case of pregnancy (Sohlberg, 2020).
