University of Groningen
Real-time positron emission tomography for range verification of particle radiotherapy
Ozoemelam, Ikechi
DOI:
10.33612/diss.133158935
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2020
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Ozoemelam, I. (2020). Real-time positron emission tomography for range verification of particle radiotherapy. University of Groningen. https://doi.org/10.33612/diss.133158935
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Real-time Positron Emission Tomography for
Range Verification of Particle Radiotherapy
© 2020 Ikechi Ozoemelam
Printed by Copy 76
PhD thesis
To obtain the degree of PhD at the
University of Groningen
on the authority of the
Rector Magnificus Prof. C. Wijminga and in
accordance with
the decision by the College of Deans.
This thesis will be defended in public on
Monday 28 September 2020 at 9:00 am
Ikechi Samuel Ozoemelam
born on January 11, 1988
in Ikorodu, Nigeria
Real-time Positron Emission Tomography for
Range Verification of Particle Radiotherapy
Supervisor
Prof. S. Brandenburg
Co-supervisor
Dr. P. Dendooven
Assessment committee
Prof. J. M. Schippers
Prof. V. Rosso
Prof. B. Poppe
Contents
1. Introduction ... 1
1.1 Rationale for particle therapy ... 1
1.2 Emerging interest in helium therapy ... 3
1.3 Aim and Outline of the Thesis ... 5
1.4 References ... 7
2.
In vivo dose delivery verification... 112.1 Uncertainties in treatment dose delivery ... 11
2.2 In vivo verification of dose delivery in particle therapy ... 12
2.2.1 Positron emission tomography ... 13
2.2.2 Prompt gamma detection ... 18
2.2.3 Iono-acoustic imaging ... 19
2.3 Conclusion ... 20
2.4 References ... 21
3. The production of positron emitters with millisecond half-life during helium
beam radiotherapy ... 293.1 Introduction ... 31
3.2 Materials and methods ... 33
3.2.1 General consideration ... 33
3.2.2 An overview of the method ... 33
3.2.3 Setup of beam irradiation and detector ... 35
3.2.4 Data analysis ... 37
3.3 Results ... 38
3.3.1 Production of positron emitters ... 38
3.3.2 Corrections for escaping positrons and photon attenuation ... 45
3.3.3 Production of PET nuclides in tissue materials and PMMA ... 47
3.3.4 Number of beam-on PET decays ... 50
3.4 Discussion ... 53
3.4.1 Benchmarking of the method ... 53
Contents
3.4.3 Beam-on PET decays ... 54
3.4.4 Feasibility of quasi-prompt range verification using short-lived nuclides . 55 3.5 Conclusion ... 56
3.6 References ... 57
3.7 Supplementary material ... 61
4. Real-Time PET imaging for range verification of helium radiotherapy ... 67
4.1 Introduction ... 69
4.2 Materials and Methods ... 71
4.2.1 Irradiation setup ... 71
4.2.2 Target and PET scanner setup ... 71
4.2.3 PET system... 73
4.2.4 Image Reconstruction ... 73
4.2.5 Reconstruction of the short-lived positron emitter contribution. ... 75
4.2.6 Detection of range shifts ... 75
4.3 Results ... 76
4.3.1 Time spectrum of activity ... 76
4.3.2 Imaging of 12N ... 77
4.3.3 Range verification using 12N ... 80
4.4 Discussion ... 85
4.5 Conclusion ... 89
4.6 References ... 90
5. Feasibility of Quasi-Prompt PET-based Range Verification in Proton Therapy
... 975.1 Introduction ... 99
5.2 Materials and Methods ... 101
5.2.1 Irradiation setup ... 101
5.2.2 Target configurations and irradiations... 102
5.2.3 PET Scanner Description ... 103
5.2.4 Scanner Sensitivity Measurement ... 105
5.2.5 Data Analysis ... 106
5.3 Results ... 109
5.3.1 Scanner sensitivity... 109
5.3.2 12N nuclide identification ... 110
5.3.3 Imaging of 12N ... 111
5.3.4 Range measurement using 12N ... 113
5.4 Discussion ... 118
5.5 Conclusion ... 122
5.6 References ... 124
6. Summary and Outlook ... 129
6.1 Introduction ... 129
6.2 Real-time imaging of short-lived positron emitters during helium beam radiotherapy ... 129
6.2.1 Production of very short-lived positron emitters for PET imaging of helium radiotherapy ... 130
6.2.2 Imaging of short-lived positron emitters ... 131
6.3 Real-time imaging of short-lived positron emitters during proton therapy .... 131
6.4 Future developments of optimal scanners. ... 132
6.5 Towards clinical implementation ... 134
6.6 References ... 135
List of Publications ... 137
Nederlandse Samenvatting ... 139
Thesis Abstract ... 145
About the Authour... 147
Chapter 1
Introduction
1.1 Rationale for particle therapy
Cancer is a leading cause of death worldwide and poses considerable public health concerns. In Europe, the change in population composition is expected to result in an increase in the cancer burden from the 3.4 million diagnosed cases recorded in 2012 to 4 million new cancer cases by 2025 (Borras et al 2016). Radiotherapy plays a role in the multidisciplinary approaches for treating cancers. It is estimated that, regardless of the region in the world, approximately 50% of cancer patients have indications for which treatment with radiation in either stand-alone treatment regimen or as a combinational therapy with surgery and/or chemotherapy is recommended (Delaney et al 2005, Barton et al 2014, Atun et al 2015).
The application of radiation for treating cancers quickly followed the discovery of X-rays by Wilhelm Roentgen in 1895 and the discovery of radium by Marie and Pierre Curie in 1898 (Connell and Hellman, 2009). Although the nascent era of radiotherapy witnessed the cure of some superficial tumours using the external beam delivery technique, treatment of deep-seated tumours was largely unsuccessful owing to limitations attributed to the low energies delivered by the available X-ray tubes and the realization that applying higher radiation doses, to achieve cure, results in damage to normal tissues. The recognition of the potential of radiation to damage normal tissues has since been captured in the principles of curative radiotherapy, whereby a tumoricidal dose of radiation is accurately conformed to the tumour volume while ensuring that there is maximal sparing of normal tissue.
Over the years, weighty advancements in treatment techniques and technology, guided by the principles of increased conformity and maximum sparing of co-irradiated tissues, have paved the road towards contemporary radiotherapy (Thariat et al 2013). Some of these advancements include technological advancements in photon (X-ray and γ-rays) production and delivery, improvements in imaging and computer-based treatment planning (Connell and Hellman, 2009). These advancements are incorporated into modern treatment techniques such as three-dimensional conformal radiation therapy (3D-CRT), intensity modulated radiotherapy (IMRT), volumetric modulated arc therapy (VMAT), helical tomotherapy and image guided radiotherapy (IGRT). They have also contributed to improved experiences for patients. The use of an IMRT technique in treating prostate cancer, for example, allows the delivery of a high dose (> 80 Gy), yielding high local control rates with minimal treatment toxicities (Spratt et al 2017). The reduction in treatment toxicities as seen in this example is attributed to a careful adherence to the dose constraints in the treatment plan and the enhanced dose conformity which limits the volumes of normal tissues exposed to high radiation dose.
1.1 Rationale for particle therapy Despite the application of these modern treatment techniques and the ensuing conformal high dose distributions, there is still a substantial volume of healthy tissue exposed to low and intermediate radiation doses, leading to the exposure of the patient to a high integral dose. The high integral dose stems from the limitations imposed by the physics of the interaction of photons, neutrons and electrons with matter. Figure 1.1 shows the dose distribution of different radiotherapy beams as a function of depth in water. The differences in the dose distribution of the particles shown highlight the various interaction mechanisms of charged particles (heavy charged particles (for e.g. protons and carbon ions) and electrons) and neutral particles (photons and neutrons). Charged particles interact via closely spaced ionization and excitation of atoms, losing energy steadily till they are brought to rest at a finite range. As the energy loss is inversely proportional to the velocity of the particles, the depth dose distribution of charged particles is characterized by a low dose in the entrance channel with a prominent dose peak, known as the Bragg peak, at a precise, energy dependent depth and almost no dose beyond the range of the particles. In contrast to heavy charged particles, the lower energy electron beams (4 – 20 MeV) typically used in radiotherapy, are brought to rest relatively close to the entrance (1 – 5 cm). Furthermore, due to the large angle scattering, no Bragg peak is seen in their depth dose distribution. Thus electrons are more suited for treatment of superficial tumours. Photons and other neutral particles such as neutrons, on the other hand, do not steadily lose energy. They travel substantial distances in between their interactions with the atoms in the materials. Single interactions with the atoms can, depending on the energy and the materials traversed, lead to complete absorption or scatter with or without significant energy loss in the medium. As the attenuation of these particles follows an exponential fall-off with depth, the depth-dose distribution features an increase in dose before a peak dose close to the entrance, where secondary electron equilibrium is established, followed by an exponential reduction of dose with depth.
For clinical therapeutic applications in deeply seated tumours, for e.g. that shown in figure 1.1 with a thickness of about 5 cm in the beam direction and located 20 cm from the entrance, charged particles offer a clear advantage. Their finite range and the Bragg peak ensures that healthy tissues along the beam path, in front and behind the tumour, are spared from a significant fraction of the total dose. Thus, charged particles offer the highest degree of conformity and enable the realization of a lower integral dose distribution (Paganetti, 2012). In addition to the Bragg peak, protons, down to a certain energy within the clinically useful energy range, and other heavier charged particles show less lateral scattering with depth in comparison to photons. Thus, heavy charged particles offer improved spatial dose selectivity which enhances the sparing of co-irradiated organs at risk (OAR). An additional boost in therapeutic effect, through a higher biological effectiveness, is observed with heavy ion (Z > 1) treatment. Particularly, there is an enhancement of the relative biological effectiveness (RBE) (Barendsen et al, 1963; Kraft, 1987 as cited in Krämer et al, 2003) and a reduction of the oxygen enhancement ratio (ratio of dose producing similar biological effect in hypoxic and oxic conditions) (Furusawa et al, 2000). The additional therapeutic enhancement due to the RBE is relevant to the extent that the RBE vs energy is peaked within the tumour volume.
Figure 1.1: Depth dose distribution of different radiation beams in water. The tumour volume considered is indicated in colour. Figure is adapted from Amaldi and Kraft (2005). Reprinted with permission.
1.2 Emerging interest in helium therapy
The first proposal to use heavy charged particles in radiation therapy dates back to 1946 and is described in Robert Wilson’s seminal paper on “Radiological use of fast protons” (Wilson, 1946). The next few years after this proposal were dedicated to understanding the biological effects of charged particles on rodents (Tobias et al, 1952 and 1954) and an upgrade of the 184-inch synchrocyclotron (Reimers et al, 1989), paving the way for the first treatments with protons in 1954 (Tobias et al 1958) and helium ions in 1955 (Cleveland et al, 1960) at the Lawrence Berkeley Laboratory (LBL). Developments in accelerator designs, resulting in the merging of LBL’s Heavy Ion Linear Accelerator (HILAC) and the Bevatron proton accelerator to form the BEVALAC, lead to the use of carbon, neon, silicon, and argon beams in clinical trials between 1975 and 1992 (Lillis-Hearne and Castro, 1995). Towards present-day treatment facilities, the history of particle therapy witnessed a transition from nuclear physics laboratories to the first dedicated therapy center at the Harvard Cyclotron laboratory in 1961 and following resolution of the technical challenges and cost, to hospital-based centers starting with the Clatterbridge Cancer center, United Kingdom in 1989 and the Loma Linda center in the United States in 1990 (PTCOG, 2020a).
Today, patients are treated with protons and carbon ion beams in 90 and 11 dedicated treatment facilities worldwide respectively (PTCOG, 2020b). In the Netherlands, three facilities are currently treating patients with protons. These facilities are the UMCG Proton Therapy Center in Groningen, HollandPTC in Delft and ZON-PTC in Maastricht. However, protons and carbon are not optimal for treatment in all circumstances, especially when there are concerns about risks to normal tissue. Protons, on the one hand, show a higher lateral scattering than heavier ions of the same range, which also exceeds that of photon beams beyond a penetration depth of 7 cm (Kraft, 2000). Carbon ions, on the other hand, despite showing reduced lateral scattering as well
1.2 Emerging interest in helium therapy as higher biological effectiveness than protons, may not represent an ideal ion for all treatments (Remmes et al, 2011). The spatially variable RBE of carbon and its associated dependencies results in an increased uncertainty in the parameters for biological optimization of therapy (Tommasino et al, 2015). In addition to the RBE related problems of carbon ions, the fragmentation tail beyond the Bragg peak distorts the much desired steep distal dose gradients. Therefore, concerns about the suitability of both protons and carbon ions have triggered a search for the optimum ion for treating cancer with enhanced capability of high dose conformation to the tumour target and sparing of OAR (Grün et al, 2015; Tommasino et al, 2015; Durante and Paganetti, 2016; Kempe et al, 2007).
This search may not result in an all-purpose optimum ion considering that the quality of particle therapy relies on the interplay of physical and biological properties of the ions as well as the configuration of the treatment field (Grün et al, 2015). This means that a particular ion with ideal physical properties may show deficiencies when considered in the light of its biological effectiveness and suitability to the treatment field configuration. The appeal of helium ions is based on a “middle-ground” advantage over the commonly used proton and carbon ion beams. From a physical perspective, helium ion beams show a smaller penumbra and less range straggling than proton beams (See e.g. Ströbele et al 2012 and Durante and Paganetti 2016). Thus, compared to proton beams, the physical properties of helium ions ensure higher conformity of dose distributions to the target (Ströbele et al 2012 and Kaplan et al 1994). Although carbon ions, due to their heavier mass, allow a still smaller penumbra than helium beams, the presence of a fragmentation tail in their depth-dose profile deteriorates the distal dose gradient of the Bragg peak (Sihver et al 1998). Since helium ions undergo less fragmentation than carbon ions (Rovituso et al 2017), they provide a good alternative for preservation of a sharp distal fall-off to negligible dose. When compared with carbon ions under biological consideration, they show a low linear energy transfer (LET) in the entrance channel and a reduced sensitivity to RBE uncertainty.
Historically, treatment with helium beams was performed only at Lawrence Berkeley Laboratory (LBL) just after the first treatments with proton beams and during the BEVALAC era (1975-1992) (Raju, 1996). In these periods, around 2000 patients were treated with helium beams for diseases including systemic disorders - controlled via pituitary irradiation, uveal melanoma, arteriovenous malformations (AVM), and large-field radiotherapy. Local control rates greater than 60%, depending on the treatment sites, were obtained for large-field radiotherapy with helium beams. Despite the impressive results obtained within this period, treatment with helium beams and other heavy ions was discontinued after the withdrawal of operational funding for the nuclear physics program at the Berkeley laboratory. Given the attractive dosimetric properties of helium ions, the interest in the utilization of helium ions in hadron therapy has recently increased again (Durante and Paganetti 2016, Tommasino et al 2015, Kempe et al 2006, Knäusl et al 2016, Grün et al 2015), with implementation planned for centres such as the Heidelberg Ion Beam Therapy Center (HIT) (Krämer et al 2016, Mairani et al 2016 and Tessonnier et al 2018). Considering the enhanced reduction of low and intermediate dose exposure of surrounding normal tissues, in silico treatment planning studies have demonstrated potential benefits from helium therapy for paediatric patients (Knäusl et al, 2016). A further in silico study using a dedicated treatment planning system by Tessonnier et al (2018) compared the dose distributions obtained for irradiations with protons and helium ions on 4 patients with brain and ocular meningiomas. The treatment plans were optimized taking into account a phenomenologically determined variable RBE approach for both ions (Mairani et al 2016 and 2017) as well as an RBE of 1.1 for protons. An evaluation of the dose volume histograms (DVH), which illustrates the volume of target structures receiving a given dose level, shows that although an almost similar planning
target volume (PTV) coverage can be obtained with both beams, helium ions, in almost all evaluated cases, show a smaller dose exposure to critical structures (Figure 1.2). Indeed, these results underscore the superior capability of helium ions to spare surrounding normal tissues.
Figure 1.2: Dose volume histograms (DVH) of the treatment plans for two patients. The DVH of the planning target volumes (PTV) and critical structures as indicated in the legend are shown. H and H(RBE=1.1) refer to the dose from protons with variable and fixed RBE respectively. The helium dose is represented by solid lines. The figure is adapted from Tessonnier et al (2018).
1.3 Aim and Outline of the Thesis
In practice, realization of the superior spatial dose selectivity of charged particles is hampered by uncertainties in the prediction of the particle range. Because these particles stop in the body, dose delivery verification techniques used in photon therapy are not applicable in particle radiotherapy. Secondary emissions resulting from the interactions of these charged particles with the tissues traversed provide surrogate signals of the particle trajectory and thus can be used to monitor the accuracy of dose delivery. One type of secondary emission is the annihilation photons. These photons are emitted during the decay of beam-induced positron emitters.
A number of investigations to assess the feasibility and in some cases clinical implementation of dose delivery verification using PET in treatments with protons (For e.g. Maccabee et al, 1969, Paans and Schippers 1993, Ferrero et al 2018), lithium (Priegnitz
1.3 Aim and Outline of the Thesis et al, 2008), carbon (For e,g. Enghardt et al 2004) and oxygen (For e.g. Bennett et al 1975, Sommerer et al, 2009) ions have been performed. Considering the growing interest in helium beam therapy and the fact that, like other ions, helium beams are also subject to range uncertainty, the monitoring of helium beam therapy can provide additional information on the accuracy of dose delivery. In contrast to other ions, there is a paucity of studies on PET monitoring of helium beam therapy. Early investigations into verification of a helium beam (Maccabee et al, 1969) show that positron emitting isotopes (for e.g. 11C (T1/2 =20.3 m), 15O (T1/2 =2.05 m)) are produced on carbon-rich materials and soft tissues, and could potentially indicate the range of the beams, provided that technical limitations of the prevalent imaging hardware could be resolved. The limitations experienced at that time were related to the unavailability of on-line detection systems which allow the detection of short-lived nuclides and reduce biological washout of the nuclides; signal deterioration by background radiation; poor detector resolution and sensitivity; absence of computing power for image reconstruction and Monte Carlo simulations; absence of cross section measurements; absence of CT imaging and absence of treatment planning systems. Several decades after this investigation, most of these limitations have received significant attention and detection systems for on-line monitoring with improved detector resolution and sensitivity (see section 2.2.1.1) and methods for suppressing background radiation (Crespo et al, 2005) have been developed. A more recent investigation into the feasibility of in-beam PET for therapeutic 3He beams (Fiedler et al, 2005) provides a quantitative estimation of the production rates of the relevant radionuclides mentioned in (Maccabee et al, 1969), and highlights significant reduction in measured activity levels, especially in oxygen-rich materials, when changing from an in-beam detection to an off-line method.
Though these studies, including those with protons, identified short-lived nuclides, for example 15O, Dendooven et al (2015 and 2019) show that positron emitting nuclides with short half-lives are produced during proton irradiation. One such short-lived positron emitter is 12N with a half-life of 11 ms. By imaging this positron emitter and other short-lived positron emitters, faster feedback on the accuracy of the treatment can be realized. The provision of feedback at a timescale comparable to about five times 12N half-lives will provide a trigger for implementation of corrective actions such as daily adaptive therapy (Albertini et al 2019) when deviations from the treatment plan are observed during imaging. This thesis is therefore aimed at investigating the production of these short-lived positron emitters during helium radiotherapy and also providing information on the feasibility of near real-time feedback on dose delivery for proton and helium beam irradiations when imaging these nuclides. Through a collaboration with Siemens, a scanner suitable for imaging of short-lived positron emitters has been realized and thus allows studies in clinical conditions.
In chapter 2, the rationale and the principles of currently investigated in vivo range verification techniques are discussed. Furthermore, the practical implementation of PET-based techniques including the imaging of short-lived positron emitters are reviewed.
In chapter 3, the yields of relevant short-lived positron emitters produced in 3He and 4He irradiations are presented to provide a first line assessment of the feasibility of imaging short-lived positron emitters for monitoring helium beam radiotherapy.
In chapter 4 and 5, the performance of near real-time range verification on the basis of imaging very short-lived positron emitters for monitoring helium and proton therapy is presented respectively. A more comprehensive description of the scanner system is provided in chapter 5.
A summary of the results and an outlook to future improvements and implementation strategies are given in chapter 6.
1.4 References
Albertini F, Matter M, Nenoff L, Zhang Y and Lomax A 2019 Online daily adaptive proton therapy BJR 93 20190594
Atun R, Jaffray D A, Barton M B, Bray F, Baumann M, Vikram B, Hanna T P, Knaul F M, Lievens Y, Lui T Y M, Milosevic M, O’Sullivan B, Rodin D L, Rosenblatt E, Van Dyk J, Yap M L, Zubizarreta E and Gospodarowicz M 2015 Expanding global access to radiotherapy Lancet Oncol 16 1153–86
Barton M B, Jacob S, Shafiq J, Wong K, Thompson S R, Hanna T P and Delaney G P 2014 Estimating the demand for radiotherapy from the evidence: A review of changes from 2003 to 2012 Radiother Oncol 112 140–4
Borras J M, Lievens Y, Barton M, Corral J, Ferlay J, Bray F and Grau C 2016 How many new cancer patients in Europe will require radiotherapy by 2025? An ESTRO-HERO analysis Radiother Oncol 119 5–11
Barendsen G W, Walter H M D, Fowler J F and Bewley D K 1963 Effects of Different Ionizing Radiations on Human Cells in Tissue Culture: III. Experiments with Cyclotron-Accelerated Alpha-Particles and Deuterons Radiat Res 18 106–19
Bennett GW, Goldberg AC, Levine GS, Guthy J, Balsamo J, and Archambeau JO 1975 Beam localization via 15O activation in proton-radiation therapy Nuclear Instruments and Methods 125 333-338
Cleveland A, Braun‐Cantilo J, Laroche G, Tobias C, Constable J, Born J, Sangalli F, Carlson R and Lawrence J H 1960 Alpha particle pituitary irradiation in metastatic carcinoma of the breast. Metabolic effects Cancer 13 190–5
Connell P and Hellman S 2009 Advances in Radiotherapy and Implications for the Next Century: A Historical Perspective Cancer Res 69 383–92
Crespo P, Barthel T, Frais-Kolbl H, Griesmayer E, Heidel K, Parodi K, Pawelke J and Enghardt W 2005 Suppression of random coincidences during in-beam PET measurements at ion beam radiotherapy facilities IEEE Trans Nucl Sci 52 980–7
Dendooven P, Buitenhuis H J T, Diblen F, Heeres P N, Biegun A K, Fiedler F, van Goethem M-J, van der Graaf E R and Brandenburg S 2015 Short-lived positron emitters in beam-on PET imaging during proton therapy Phys. Med. Biol. 60 8923–47
Dendooven P, Buitenhuis H J T, Diblen F, Heeres P N, Biegun A K, Fiedler F, Goethem M-J van, Graaf E R van der and Brandenburg S 2019 Corrigendum: Short-lived positron emitters in beam-on PET imaging during proton therapy (2015 Phys. Med. Biol. 60 8923) Phys. Med. Biol. 64 129501
Durante M and Paganetti H 2016 Nuclear physics in particle therapy: a review Rep. Prog. Phys. 79 096702
Delaney G, Jacob S, Featherstone C and Barton M 2005 The role of radiotherapy in cancer treatment Cancer 104 1129–37
Enghardt, W., Crespo, P., Fiedler, F., Hinz, R., Parodi, K., Pawelke, J. and Pönisch, F. (2004). Charged hadron tumour therapy monitoring by means of PET. Nucl Instrum Methods Phys Res A 525 284-288.
Fiedler F, Crespo P, Parodi K, Sellesk M and Enghardt W 2005 The feasibility of in-beam PET for therapeutic beams of /sup 3/He IEEE Nucl Sci Symp Conf Rec, 2005 vol 5 pp 2757–60 Furusawa Y, Fukutsu K, Aoki M, Itsukaichi H, Eguchi-Kasai K, Ohara H, Yatagai F, Kanai T and
Ando K 2000 Inactivation of Aerobic and Hypoxic Cells from Three Different Cell Lines by Accelerated 3He-, 12C- and 20Ne-Ion Beams rare 154 485–96
Ferrero V, Fiorina E, Morrocchi M, Pennazio F, Baroni G, Battistoni G, Belcari N, Camarlinghi N, Ciocca M, Guerra A D, Donetti M, Giordanengo S, Giraudo G, Patera V, Peroni C, Rivetti A, Rolo M D da R, Rossi S, Rosso V, Sportelli G, Tampellini S, Valvo F, Wheadon R, Cerello P and Bisogni M G 2018 Online proton therapy monitoring: clinical test of a Silicon-photodetector-based in-beam PET Sci Rep 8 4100
1.4 References Grün R, Friedrich T, Krämer M, Zink K, Durante M, Engenhart-Cabillic R and Scholz M 2015 Assessment of potential advantages of relevant ions for particle therapy: A model based study Med Phys 42 1037–47
Kraft G 2000 Tumor therapy with heavy charged particles Prog Part Nucl Phys 45 S473–544
Krämer M, Scifoni E, Schuy C, Rovituso M, Tinganelli W, Maier A, Kaderka R, Kraft-Weyrather W, Brons S and Tessonnier T 2016 Helium ions for radiotherapy? Physical and biological verifications of a novel treatment modality Med Phys 43 1995–2004
Krämer M, Weyrather W K and Scholz M 2003 The increased biological effectiveness of heavy charged particles: from radiobiology to treatment planning Technol. Cancer Res. Treat. 2 427–36 Kempe J, Gudowska I and Brahme A 2007 Depth absorbed dose and LET distributions of
therapeutic 1H, 4He, 7Li, and 12C beams Med Phys 34 183
Knäusl B, Fuchs H, Dieckmann K and Georg D 2016 Can particle beam therapy be improved using helium ions? – a planning study focusing on pediatric patients Acta Oncol 55 751–9 Kaplan I D, Castro J R and Phillips T L 1994 Helium charged particle radiotherapy for
meningioma: experience at UCLBL. University of California Lawrence Berkeley Laboratory Int. J. Radiat. Oncol. Biol. Phys. 28 257–61
Lillis-Hearne, P. and Castro, J., 1995. Indications for Heavy Ions - Lessons from Berkeley. In: U. Linz, ed., Ion Beams in Tumour Therapy, 1st ed. Weinheim: Chapman & Hall GmbH, pp.133 - 141. Mairani A, Magro G, Dokic I, Valle S M, Tessonnier T, Galm R, Ciocca M, Parodi K, Ferrari A, Jäkel O, Haberer T, Pedroni P and Böhlen T 2016 Data-driven RBE parameterization for helium ion beams Phys. Med. Biol. 61 888–905
Maccabee, H., Madhvanath, U. and Raju, M. (1969). Tissue activation studies with alpha-particle beams. Phys. Med. Biol. 14 213-224.
Mairani A, Dokic I, Magro G, Tessonnier T, Bauer J, Böhlen T, Ciocca M, Ferrari A, Sala P R, Jäkel O, Debus J, Haberer T, Abdollahi A and Parodi K 2017 A phenomenological relative biological effectiveness approach for proton therapy based on an improved description of the mixed radiation field Phys Med Biol 62 1378–95
Paans A M J and Schippers J M 1993 Proton therapy in combination with PET as monitor: a feasibility study IEEE Trans Nucl Sci 40 1041–4
PTCOG, 2020a. PTCOG - Facilities in Operation. [online] Ptcog.ch. Available at:
https://www.ptcog.ch/images/patientstatistics/Patientstatistics-updateDec2018.pdf [Accessed 15 June 2020].
PTCOG, 2020b. PTCOG - Facilities in Operation. [online] Ptcog.ch. Available at:
https://www.ptcog.ch/index.php/facilities-in-operation [Accessed 9 March 2020].
Priegnitz M, Möckel D, Parodi K, Sommerer F, Fiedler F and Enghardt W 2008 In-beam PET measurement of 7Li3+ irradiation induced beta+-activity Phys Med Biol 53 4443–53
Raju M R 1996 Particle Radiotherapy: Historical Developments and Current Status Radiat Res 145 391
Rovituso M, Schuy C, Weber U, Brons S, Cortés-Giraldo M A, La Tessa C, Piasetzky E, Izraeli D, Schardt D, Toppi M, Scifoni E, Krämer M and Durante M 2017 Fragmentation of 120 and 200 MeV u −14 He ions in water and PMMA targets Phys. Med. Biol. 62 1310–26
Reimers R, Haley J and Hampton G 1989 184-inch synchrocyclotron decommissioning Proceedings of the 1989 IEEE Particle Accelerator Conference, ’Accelerator Science and Technology Proceedings of the 1989 IEEE Particle Accelerator Conference, ’Accelerator Science and Technology pp 1605–7 vol.3
Remmes N B, Herman M G and Kruse J 2012 Optimizing Normal Tissue Sparing in Ion Therapy Using Calculated Isoeffective Dose for Ion Selection Int. J. Radiat. Oncol. Biol. Phys 83 756–62 Tommasino F, Scifoni E and Durante M 2015 New Ions for Therapy Int. J. Part. Ther. 2 428–38
Tessonnier T, Mairani A, Chen W., Sala, P., Cerutti, F., Ferrari, A., Haberer, T., Debus, J., & Parodi, K. 2018 Proton and helium ion radiotherapy for meningioma tumors: a Monte Carlo-based treatment planning comparison. Radiat Oncol. 13 2
Thariat J, Hannoun-Levi J-M, Sun Myint A, Vuong T and Gérard J-P 2013 Past, present, and future of radiotherapy for the benefit of patients Nat. Rev. Clin. Oncol. 10 52–60
Tobias C A, Anger H O and Lawrence J H 1952 Radiological use of high energy deuterons and alpha particles Am J Roentgenol Radium Ther Nucl Med 67 1–27
Tobias C A, Lawrence H, Born L, McCoMBS R K, Roberts J E, Anger H O and Low-BEER B V A 1958 Pituitary Irradiation with High-Energy Proton Beams a Preliminary Report Cancer Res 18 20
Tobias C A, Van Dyke D C, Simpson M E, Anger H O, Huff R L and Koneff A 1954 Irradiation of the pituitary of the rat with high energy deuterons Am J Roentgenol Radium Ther Nucl Med 72 1–21
Ströbele J, Schreiner T, Fuchs H and Georg D 2012 Comparison of basic features of proton and helium ion pencil beams in water using GATE Z Med Phys 22 170–8
Spratt D E, Pei X, Yamada J, Kollmeier M A, Cox B and Zelefsky M J 2013 Long-term Survival and Toxicity in Patients Treated with High-Dose Intensity Modulated Radiation Therapy for Localized Prostate Cancer Int. J. Radiat. Oncol. Biol. Phys 85 686–92
Sommerer F, Cerutti F, Parodi K, Ferrari A, Enghardt W and Aiginger H 2009 In-beam PET monitoring of mono-energetic 16O and 12C beams: experiments and FLUKA simulations for homogeneous targets Phys Med Biol 54 3979–96
Sihver L, Schardt D and Kanai T 1998 Depth-dose distributions of high-energy carbon, oxygen and neon beams in water Igaku Butsuri.18 1–21
Chapter 2
In vivo dose delivery verification
Particle therapy promises to deliver conformal dose to tumour volumes while sparing surrounding healthy tissues and organs-at-risk. The uncertainties in accurately localizing the Bragg peak in the body potentially compromise the realization of this promise. In this chapter, the causes of uncertainties in the dose delivery and strategies to mitigate range uncertainties including in vivo range verification are reviewed.
2.1 Uncertainties in treatment dose delivery
Regardless of the choice of radiation used, charged particles all have a distinctive feature of a finite range in matter which enables a more or less accurate alignment of the Bragg peak in the tumour volume. However, uncertainties in the ion range, due to factors including patient immobilization and setup errors, patient and organ motion, anatomical changes (tumour shrinkage, fluid and air filling of cavities and weight loss), imaging artifacts, uncertainties in the conversion of CT numbers to stopping power ratios (SPR) and dose calculations can lead to deviations of the delivered dose distribution with respect to the intended one, causing inadvertent exposure of organs at risk and/or under-dosage of the tumour volume (Goitein 1985, Lomax 2008 and Paganetti 2012). In current clinical practice, a combination of several strategies has been implemented to mitigate the impact of these uncertainties. The strategies comprise the enhancement of the robustness of the beam to under- or overshoot; and reduction of the uncertainties by imaging the patient more accurately. The common strategy involves the adherence to some basic rules in the design of treatment plans. These basic rules govern the selection of beam directions such that less heterogeneity is traversed by the beam and ensuring that the beam never stops in front of organs at risk (OARs) located close to the tumor. This approach is employed, for example, in the treatment of prostate tumours, where a longer particle track via a lateral beams rather than a shorter anterior beam is used to improve sparing of the rectum. A further mitigation strategy involves the addition of an extra safety margin, which accounts for range uncertainties, to the distal margins of the planning target volume (PTV) (van Herk et al 2000, Albertini et al 2011, Paganetti 2012, Lomax 2019). Depending on the treatment facility, a range margin composed of 2.5-3.5% of the prescribed range plus 1-3 mm is used (Paganetti 2012). It is expected that under conditions of range over- or undershoot, the clinical target volume (CTV) still receives the prescribed dose as the CTV remains completely enclosed by the PTV. Although the PTV concept works well for photon therapy, it does not sufficienctly act as a buffer against dose inhomogeneity in intensity modulated proton therapy (IMPT) due to the degradation of the static dose distribution in non-nominal scenarios. Using such margins may result in hot and cold spots within the CTV in uncertain scenarios due to range undershoot and overshoot respectively. Furthermore, the additional margins result in a deliberate delivery of the prescribed dose to a considerable volume of normal tissues. For instance, a range margin of 3.5% + 1 mm is used in the MGH proton therapy center representing a margin of 8
2.1 Uncertainties in treatment dose delivery mm for a 20 cm range. Recently, robust planning techniques, which incorporate the probable uncertainty scenarios (range uncertainties and setup errors) into the treatment plan optimization algorithms have been implemented in treatment planning systems (see review by Unkelbach et al 2018 and articles by Unkelbach et al 2007 and 2009, Pflugfelder et al 2008, Fredriksson et al 2011). The use of such robust optimization have been shown to allow a reduction in dose to healthy tissues and organs at risk with good coverage of the target. The adoption of these robust planning techniques, however, still leads to the delivery of extra dose to normal tissues, compared to plan optimization on the CTV, as considerable trade-offs between target coverage and dose to surrounding structures are implicit to their implementation. Consequently, such techniques may compromise the reduction of radiation-induced side effects which is the main rationale for proton therapy. A more fundamental solution to reducing range uncertainties is to improve the accuracy of the Stopping Power Ratio (SPR) required by the treatment planning system for calculation of the proton range and water equivalent path length (WEPL). The SPRs are commonly estimated using the stoichiometric calibration method to obtain a CT numbers-to-SPR calibration curve (Schneider et al 1996, Schneider et al 2000). The disadvantage of this technique is that there is a degeneracy between the CT numbers and the estimated SPRs. This implies that tissues with the same CT numbers could have different SPRs and vice versa. The degeneracy problem is attributed to the incapability of the CT numbers, obtained by using a single X-ray spectrum for imaging, to capture the variability of elemental composition in human tissues (Yang et al, 2010). The determination of the SPR depends on elemental properties such as the electron density ratios and effective atomic number or the mean excitation energy. Dual-energy CT, which involves CT imaging with two different energy spectra, has been shown to enable simultaneous retrieval of two important tissue properties – relative electron densities and effective atomic number – and thus improve the accuracy of the estimated SPRs to about 1% (Hünemohr et al 2013, Bourque et al 2014, Hudobivnik et al 2016, Möhler et al 2016, Han et al 2016, Taasti et al 2016, Lalonde et al 2017, Saito and Sagara 2017, van Abbema 2017, Almeida et al 2018). A more ideal approach to estimating the SPR is to directly perform computed tomographic imaging of the patients with ion beams (Cormack 1963, Huesman et al 1975). Recent experimental studies on phantoms have demonstrated the capabilities of ion CT to determine SPR with accuracies of 1.6% and below 1% for proton (Esposito et al 2018) and helium (Volz et al 2018) ions respectively.
2.2 In vivo verification of dose delivery in particle therapy
The essentially 3D nature of the dose deposition by a particle beam makes in vivo treatment verification more important than for highly penetrating MV X-ray photon beams, whose dose deposition is basically 2D in nature. With the increasing complexity of treatment techniques employed in photon radiotherapy, the use of an Electronic Portal Imaging Device (EPID) for in vivo treatment verification has gained considerable traction (Meertens et al 1985 and review by van Elmpt et al 2008, McCurdy et al 2017, Mijnheer 2019). While these devices detect the transmitted MV X-rays, the stopping of the particle beams in the patient precludes the use of this method in particle beam radiotherapy. Thus, a number of in vivo range verification techniques for particle beam radiotherapy, aiming at verifying the ion range with a precision of about 2 mm (Pausch et al 2020) and thereby reducing the required safety margins and also the range uncertainties considered in robust optimization, have been proposed, (See reviews by Parodi and Polf 2018, Parodi 2011, Knopf and Lomax 2013). The reduction of the safety margins, when implementing in vivo verification techniques, complemented by robust optimization potentially results in a reduction of normal tissue complications. These techniques rely mostly on the detection
of secondary emissions resulting from the interaction of the particle beam with the tissues.
2.2.1 Positron emission tomography
Traditionally, positron emission tomography (PET) is a molecular imaging modality used in nuclear medicine for the detection of activity distributions of radioactive tracers injected into the bloodstream of patients. These tracers are tagged with radionuclides, which decay by the emission of positrons. Subsequent interactions of these positrons with atomic electrons result in their annihilation, with opposite emission of a pair of annihilation photons (energy = 511 keV). The coincident detection of these photon pairs, as shown in figure 2.3, allows the characterization (quantitative and spatial) of the activity in the patient.
Figure 2.1: Principles of PET imaging. The annihilation of emitted positrons, β+, results in the production of a pair of oppositely directed photons which are detected in coincidence by a pair of detectors in the PET detector ring.
In a non-conventional approach, PET imaging can be used to monitor the range of therapeutic charged particle beams (Maccabee et al 1969, Paans and Schippers 1993, Nishio et al 2008, Enghardt et al 2004 among others; review papers by Studenski and Xiao 2010, Fiedler et al 2012, Zhu and El Fakhri 2013, Parodi and Polf 2018, Parodi 2019). As particles traverse the body, they undergo nuclear interactions with the target nuclei. Most product nuclides, such as 11C, 15O, 13N, and 38gK, are proton-rich and decay by emission
of positrons.
These interaction processes produce a positron activity along the trajectory of the particle, which, depending on the particle type, drops sharply to zero or with some activity tails. Figure 2.4 shows 1-dimensional profiles of the dose and induced activity for protons and heavier ions in polymethylmethacrylate (PMMA). Due to differences between the dose deposition and nuclear interaction processes, the activity distribution does not directly correspond to the dose distribution, though a closer correlation is obtained with projectile fragmentation. Similarly, the activity fall-off does not correspond to the dose fall-off and doesn’t explicitly indicate the range of the incident beam. It has, however, been shown that the 50% activity fall-off point, depending on the specific tissue composition, mass density and cross sections for the PET isotope production channels,
2.2 In vivo verification of dose delivery in particle therapy lies a few millimetres before the range of the primary particle and is correlated to the beam range (Oelfke et al, 1996; Parodi et al, 2005).
Figure 2.4: Measured activity and irradiation dose profiles in PMMA for protons, 3He, 7Li, 12C and 16O respectively. The energy of the ions was selected to have the same range of 80 mm in PMMA. Figure is adapted from Fiedler et al (2012). Reprinted with permission.
For range verification, a comparison of the measured activity distribution with a reference distribution is required for assessment of any range deviation. There are a number of approaches to obtain the reference distributions. One category of approaches creates the reference using the treatment plan: Monte Carlo simulations of the PET activity (Parodi et al 2002 and 2007, Seravalli et al 2012), convolutions of the calculated dose with filter functions (Parodi and Bortfeld 2006, Frey et al 2014) and scaling of measured activity yield to the treatment plan (Miyatake et al 2011 and Priegnitz et al 2012, Helmbrecht et al 2016). A second approach uses the first treatment day PET measurement (Nishio et al, 2010, Ferrero et al 2018) for comparison with subsequent fractions. The drawback of this approach is that systematic deviations present on the first day of treatment are not detected. Most investigations and clinical applications have adopted the Monte Carlo prediction approach. With a Monte Carlo simulation, the activity distribution as measured is predicted on the basis of the cross section for production channels of the relevant positron emitters, the elemental composition of the patient tissues, the charged particle fluence, influence of biological washout on activity profiles, irradiation/data acquisition time course and details of the PET scanner system. Activity profiles predicted using the internal nuclear interaction models implemented in most Monte Carlo codes shown significant deviations from those obtained using experimental cross sections (Parodi et al 2002). Thus, the estimation of activity profiles
has to be based on experimentally measured cross section data and fine-tuned cross-section data fitted to measured activity profiles (Bauer et al 2013).
2.2.1.1 PET acquisition strategies
To provide feedback on the accuracy of the predicted range of the incident beam, the PET isotopes produced are detected by imaging detectors positioned around the patient. Different strategies have been investigated for acquisition of the PET data (Zhu and El Fakhri 2013). Figure 2.5 is a representation of the various strategies for implementation of PET-based range verification. In the off-line strategy (Parodi et al 2007, Nishio et al 2008, Knopf et al 2011, Nischwitz et al 2015), the patient is transferred to a commercial full ring PET/CT scanner located outside the treatment room. This approach offers the advantages of being relatively less expensive to implement, provides complete angular coverage and allows co-registration of PET and CT information. However, due to the long interval prior to data acquisition, the radioactive decay is dominated by the longest-lived beam-induced nuclide, 11C. Given the long half-life of 11C, rather long data acquisition times to collect sufficient statistics are needed. Furthermore, the activity distribution is deteriorated due to the metabolic washout of the induced activity from the irradiated area, thus compromising the accuracy of range verification.
Figure 2.5: Three implementation strategies for in vivo PET-based range verification. (a) In-beam PET with PET scanner installed in the treatment position (b) In-room PET with PET scanner installed in the treatment room and patient couch rotated into the PET scanner (c) Offline PET using a PET scanner installed nearby the treatment room with patients moved after dose delivery. The figure is taken from Zhu and El Fakhri (2013)
To reduce the effects associated with radioactive decay and biological washout of the signal, the acquisition of the PET activity within the treatment room (In-room) has been investigated (Zhu et al 2011). In this strategy, a full-ring neuro-PET scanner was moved into the room after patient irradiation, with the patient couch sliding into the scanner for imaging. Despite the improvements in the detectable activity while using this scanner, sub-optimal information was obtained due to the absence of CT imaging capability, necessary for co-registration of the PET image with the planning CT image and attenuation correction. In addition, the PET scanner field of view was only compatible with imaging smaller body parts and children.
In the in-beam strategy (Enghardt et al 2004, Ferrero et al 2018), the patient remains immobilized in the treatment position and PET images are acquired during the irradiation and/or a short time afterwards. This strategy provides maximum counting statistics - especially from short-lived radioisotopes, fastest feedback for quality assurance and is less susceptible to biological washout of the signal. Data for in-beam PET can be
2.2 In vivo verification of dose delivery in particle therapy collected from various time points during and/or a short time after irradiation. These periods include the beam-on period (i.e during an irradiation spill or proton bunch), beam-off (i.e. during the pauses of the irradiation) and shortly after irradiation. The in-beam implementation requires the integration of the PET scanner with the in-beam delivery nozzle, which imposes geometrical limitations on the angular coverage of the scanner since the setup must guarantee unrestricted access to the patient and beam delivery nozzle. A drawback associated with the beam-on implementation is the addition of random signals to the true prompt signals, thus reducing the signal-to-background of the PET image. The additional random signals are prompt gamma photons correlated with the particle bunch. As a result of this signal deterioration, in the absence of corrections, beam-on acquisition in a synchrotron facility has been limited to the pauses in-between the spills. In a clinical cyclotron facility, with a continuous beam delivery, online acquisition has been limited to the period immediately after irradiation. With the development of random suppression and pile-up rejection algorithms using an anti-coincidence filters to subtract events correlated with the beam micro bunch (Crespo et al 2005, Helmbrecht et al 2017, Buitenhuis et al 2017), useful data can be acquired during the beam-on period.
2.2.1.2 In-beam PET imaging of short-lived positron emitters
In an in-beam PET measurement, a range of nuclides, with half-lives up to about 20 minutes, will contribute to the measured PET activity. Provision of PET-based fast feedback, on a sub-second timescale, can only be obtained through imaging of very short-lived positron emitters. The provision of such prompt feedback will provide the necessary trigger for implementation and verification of strategies, such as daily adaptive proton therapy (Albertini et al, 2020), to correct deviations from the treatment plan. Beyond conventional fractionated proton therapy, the implementation of hypo-fractionation schemes (Grewal et al 2019) and high dose rate treatment techniques such as FLASH irradiations (Favaudon et al 2014, Vozenin et al 2019) with small safety margins might benefit from fast feedback at short/ultrashort timescale using a probe fraction of the full dose to boost confidence in the treatment delivery. The most important of the very short-lived nuclides for proton therapy, 12N (T1/2 = 11 ms), has been shown to dominate the beam-induced PET decay in carbon-rich tissues at such short/ultrashort time scales and even up to 70 s into the irradiation (Dendooven et al 2015 and Dendooven et al 2019). A proof-of-principle study on the use of this important nuclide for near real-time proton range verification using a small PET scanner (two 6.5 × 6.5 cm2 modules) shows that an accuracy in the range measurement of 3 mm is obtained for the detection of 4000 12N PET counts during irradiations with 2.5 × 1010 protons delivered over a 120 s period (Buitenhuis et al 2017).
2.2.1.3 Instrumentation for In-Beam PET acquisition
The requirements for beam delivery nozzle access and unrestricted patient positioning in the treatment field means that a closed-ring geometrical configuration, as used in conventional PET systems, is not feasible. A configuration employing dual-head PET scanners, with an opening for the beam and patient table, was implemented as solution to this geometrical constraint at patient treatment facilities at GSI Darmstadt for monitoring of carbon ion therapy (Enghardt et al, 1999) and the National Cancer Centre (NCC) of Kashiwa, Japan (Nishio et al, 2010) for monitoring proton beams. The detector technology of the GSI PET system is composed of standard Siemens Bismuth Germanate (BGO) block detectors read-out by PMTs. For the NCC PET system, a Hamamatsu custom-built system comprised of BGO arrays coupled to multi-anode PMTs was used. Several aspects of the PET instrumentation used in these studies, such as
sensitivity and better detectors, leading to improved imaging performance have been introduced.
Considering the gains accruable from in-beam detection, improvements in the detector technology and acquisition strategies are actively investigated. Crespo et al (2007) and Surti et al (2011) have demonstrated how the inclusion of Time-of-Flight (ToF) information in the reconstruction of the coincidence event mitigates the limited angle artefacts and increases the signal-to-background ratio. Furthermore, such an approach, in principle, allows real time and direct image reconstruction by assigning the event location to a point along the line of response between the two detectors fired in coincidence, using the arrival time difference of the detected photons. The relationship between the spatial localization uncertainty, 𝛥𝑥, and the detector timing resolution, 𝛥𝑡, is given as:
𝛥𝑥 = 𝑐
𝛥𝑡
2
(1)
where 𝑐 is the speed of light. Due to the detector system timing resolution, precise localization of the event to a point along the LOR is not possible. Thus, events are distributed to voxels within a spatial uncertainty, which corresponds to the timing resolution of the PET system (Surti and Karp, 2016). Consequently, improvements in the timing resolution would enable more precise localization and full utilization of the benefits of TOF-PET. To achieve this improvement of the scintillator material and photon detection technology is required. Currently, major vendors of PET systems utilize lutetium-based scintillators coupled to either photomultiplier tubes (PMT) or silicon photomultipliers (SiPM), to achieve timing resolutions down to 400-550 ps and 200-400 ps respectively. Further improvements in the imaging performance are envisaged if TOF-PET detectors with a timing resolution below 200 ps (FWHM) become commercially available, though no significant change below 33 ps (FWHM) is expected due to the spatial resolution (5 mm) of PET systems (Crespo et al, 2007). The performance of such crystal-based TOF-PET systems using SiPMs has been investigated for online verification of the beam range (Cambraia Lopes et al 2016, Buitenhuis et al 2017, Diblen et al 2017 and Ferrero et al 2018). Although the studies of Cambraia Lopes et al (2016), Buitenhuis et al 2017 and Ferrero et al (2018) demonstrate the capabilities of their SiPM based In-beam PET systems to acquire good data, the study by Diblen et al (2017) shows that SiPMs installed in In-beam configurations are rather susceptible to radiation damage. The damage results in an increase in the dark count rate (DCR); a review of radiation damage of SiPMs is discussed by Garutti and Musienko (2019). Diblen et al (2017) estimate that successful in-beam operation of the detectors beyond a few weeks (irradiations with about 5 ×1013 protons) is limited by the large DCR. In addition to crystal-based systems, TOF-PET systems based on resistive-plate chambers have been proposed and evaluated for their performance (Couceiro et al, 2014). However, a recent comparison with crystal based systems shows that the low sensitivity of RPC systems poses a limitation to their performance in hadron therapy (Torres-Espallardo, 2015).
Besides the TOF technique, other design features aimed at improving the image quality at low count statistics as obtained in in-beam detection of induced PET activity have been investigated. With their prototype dual panel PET scanner of block detectors, enabled with Depth-of-Interaction (DOI) functionality and read out by SiPM, Shao et al (2014a and 2014b) demonstrated the feasibility of measuring the activity-range from full-tomographic images acquired through rotation of their prototype design around a phantom. The DOI functionality enhances the spatial resolution by eliminating the mis-positioning errors due to the penetration of the scanner crystals by obliquely directed gamma-rays emitted from the periphery of the Field of View (FoV). In practice, the tomographic acquisition as envisioned by the authors is inconsistent with typical
2.2 In vivo verification of dose delivery in particle therapy irradiation durations. In another application of DOI detectors for online measurement of PET activity, a slanted, closed-ring, Gadolinium Orthosilicate (GSO)-based PET detector system, with an opening for the beam portal and unrestricted access to the patient, is being developed and has been evaluated for its performance in non-therapeutic conditions (Yoshida et al, 2013) and for on-line heavy-ion therapy using Monte Carlo simulations with patient data (Tashima et al, 2015). Good imaging performance for In-beam PET was demonstrated using this prototype design. Additionally, a comparable quality image was obtained in a shorter image acquisition time than for an offline measurement using a full-ring PET system.
2.2.2 Prompt gamma detection
Secondary prompt gamma photons are energetic γ emissions by excited nuclei formed during nuclear interactions between the ion beam and the irradiated tissues. An overview on range verification using prompt gamma rays is presented in Krimmer et al (2018) and Pausch et al (2020). The prompt gamma rays are emitted within a short time after the production of the excited nucleus, typically less than 1 ns, and as such are useful for retrieval of real time feedback on the range. Due to the high energy of these photons (up to 6.5 MeV), imaging of the spatial distribution of their origin is far from trivial. Dedicated detectors and collimators are required to accommodate the high energy of the gamma photons and to exclude the non-trivial contribution of background secondary radiation such as neutrons. Following the first proposal at the 39th Particle Therapy Co-operative Group (PTCOG) meeting in 2003 (Stichelbaut and Jongen 2003), different detection approaches have been proposed. These approaches are categorized into imaging and non-imaging systems.
2.2.2.1 Imaging systems
The imaging systems employ either collimators or the directional capabilities of Compton cameras. For collimated systems, a feasibility study on the use of a slit and a single detector, for monitoring irradiations with 100 – 200 MeV protons by Min et al (2006) showed a good correlation of the prompt gamma ray profile with the distal fall-off region of the dose. Follow-up designs of the slit concepts have been proposed for collimated prompt gamma imaging (Seok et al 2006, Smeets et al 2012 and Pinto et al 2014), with the most advanced design now realized in the knife-edge slit camera (Smeets et al 2012 and Perali et al 2014). A precision in shift retrieval of 4 mm (2𝜎) when irradiating PMMA targets with 0.5 × 108, 1.4 × 108 and 3.4 × 108 protons at 100, 160 and 230 MeV, respectively, has been shown for the knife-edge system (Perali et al 2014). The system is the only prompt gamma ray detection system that has been so far tested during clinical irradiation, both with passive scattering (Richter et al 2016) and active pencil beam scanning (Xie et al 2017) delivery. Although knife edge systems have the potential to detect local range shifts down to 1–2 mm, the massive size of the collimators poses a challenge to the reproducibility of camera positioning and clinical integration (Nenoff et al 2017).
As an alternative to collimation, the direction sensitivity of Compton cameras can be utilized for range verification. Compton imaging exploits the kinematics of Compton scattering to reconstruct the vertex of the photon production. The cameras are composed of two or more layers of detectors within which the photons interact by Compton scattering or, if two layers of detectors are used, by Compton scattering and total absorption. Using these multiple interaction points and the deposited energy in the detector, the emission point can be restricted to a cone; superposition of multiple cones enables the retrieval of the prompt gamma source position. A number of Compton camera prototypes are either in development or still being constructed (Krimmer et al
2015, Thirolf et al 2016, Llosá et al 2016, Golnik et al 2016, Draeger et al 2018). Despite the capability to detect range shifts as small as 2 mm using 6.3 × 108 protons in one of the prototypes (Draeger et al 2018), additional investigations, prior to full clinical deployment, still need to be performed to resolve some challenges of the technique. Particularly, these investigations need to address the following: technical complexity of the detectors; reducing the cost of the electronics; handling of the huge radiation load during dose delivery; and enhancement of the fraction of ‘valid’ events imposed by the coincidence and event selection criteria (Pausch et al (2020))
2.2.2.2 Non-imaging systems
Non-imaging systems aim to minimize the complexity of detection systems by using direct data acquisition and simple detectors, and a potential reduction of the footprint, in some cases, of the system in the treatment room. Three such systems – prompt gamma-ray spectroscopy (Verburg and Seco 2014, Hueso-González et al 2018, Dal Bello et al 2019), prompt gamma-ray timing (Golnik et al 2014) and prompt gamma-ray peak integration (Krimmer et al 2017) – utilizing distinct features of prompt gamma events have been proposed.
The prompt gamma-ray spectroscopy technique utilizes the dependence of the intensity of prompt gamma energy lines on the proton energy or residual range to determine the absolute proton range and material composition (Verburg and Seco 2014). A recent study with a full-scale clinical prototype of the system demonstrates the measurement of absolute range with a precision of 1.1 mm (2𝜎) when aggregating pencil-beam spots within a cylindrical region of 10 mm radius and 10 mm depth, i.e. equivalent to 1.6 ×109 protons per aggregated spot (Hueso-González et al 2018).
The prompt gamma-ray timing technique relies on the sensitivity of the proton transit time in a medium on the proton range (Golnik et al 2014). The transit time is determined using a Time-of-Flight (ToF) measurement with the reference time point for the measurement obtained preferably from the accelerator radiofrequency signal. Statistical moments of the ToF spectrum such as the mean and the standard deviation are used to determine the range shifts. Due to drift of the magnetic field in the accelerator leading to instability of the beam arrival times with respect to the radiofrequency signal, alternative signals from a proton bunch monitor has been investigated for provision of more stable time reference (Petzoldt et al 2016). A test with simple phantoms under clinical conditions has shown the capability of the system composed of 8 timing detectors to detect range shifts for irradiations of 3.2 ×109 162 MeV protons with a precision of 2.7 mm and 4.9 mm when using the ToF mean and width as range shift metric (Werner et al 2019).
In contrast to the prompt gamma-ray timing technique, the prompt gamma-ray peak integration technique uses the differences in the integral of the prompt gamma-ray TOF spectrum to detect range deviations (Krimmer et al 2017, Hueso-González and Bortfeld 2020). In the study of Krimmer et al (2017), the focus on the peak of the TOF spectrum ensures that gamma photons originating from the body are discriminated from those arising from the beam line. Using a detector covering a solid angle of 25 msr (3” – 4” diameter installed 50 cm from the beam axis), the authors demonstrated the possibility to detect a range change of 3 mm when irradiating with 108 protons. Hueso-González and Bortfeld (2020) propose to develop a low-cost compact system mounted coaxially with the beam direction and capable of range verification with a precision of 1 mm
2.2.3 Iono-acoustic imaging
A promising technique for range verification, relying on emissions other than gamma photons, is iono-acoustics. A review of the technique is presented in Hickling et al (2018).
