3D prompt gamma imaging for proton beam range verification

An experimental validation of a filtering approach for prompt gamma prediction in a research proton treatment planning system

Objective.Prompt gamma (PG) radiation generated from nuclear reactions between protons and tissue nuclei can be employed for range verification in proton therapy. A typical clinical workflow for PG range verification compares the detected PG profile with a predicted one. Recently, a novel analytical PG prediction algorithm based on the so-called filtering formalism has been proposed and implemented in a research version of RayStation (RaySearch Laboratories AB), which is a widely adopted treatment planning system. This work validates the performance of the filtering PG prediction approach.Approach.The said algorithm is validated against experimental data and benchmarked with another well-established PG prediction algorithm implemented in a MATLAB-based software REGGUI. Furthermore, a new workflow based on several PG profile quality criteria and analytical methods is proposed for data selection. The workflow also calculates sensitivity and specificity information, which can help practitioners to decide on irradiation course interruption during treatment and monitor spot selection at the treatment planning stage. With the proposed workflow, the comparison can be performed on a limited number of selected high-quality irradiation spots without neighbouring-spot aggregation.Main results.The mean shifts between the experimental data and the predicted PG detection (PGD) profiles (ΔPGD) by the two algorithms are estimated to be1.5±2.1mm and-0.6±2.2mm for the filtering and REGGUI prediction methods, respectively. The ΔPGD difference between two algorithms is observed to be consistent with the beam model difference within uncertainty. However, the filtering approach requires a much shorter computation time compared to the REGGUI approach.Significance.The novel filtering approach is successfully validated against experimental data and another widely used PG prediction algorithm. The workflow designed in this work selects spots with high-quality PGD shift calculation results, and performs sensitivity and specificity analyses to assist clinical decisions.

Patient-specific deep learning for 3D protoacoustic image reconstruction and dose verification in proton therapy

BACKGROUND

Protoacoustic (PA) imaging has the potential to provide real-time 3D dose verification of proton therapy. However, PA images are susceptible to severe distortion due to limited angle acquisition. Our previous studies showed the potential of using deep learning to enhance PA images. As the model was trained using a limited number of patients' data, its efficacy was limited when applied to individual patients.

PURPOSE

In this study, we developed a patient-specific deep learning method for protoacoustic imaging to improve the reconstruction quality of protoacoustic imaging and the accuracy of dose verification for individual patients.

METHODS

Our method consists of two stages: in the first stage, a group model is trained from a diverse training set containing all patients, where a novel deep learning network is employed to directly reconstruct the initial pressure maps from the radiofrequency (RF) signals; in the second stage, we apply transfer learning on the pre-trained group model using patient-specific dataset derived from a novel data augmentation method to tune it into a patient-specific model. Raw PA signals were simulated based on computed tomography (CT) images and the pressure map derived from the planned dose. The reconstructed PA images were evaluated against the ground truth by using the root mean squared errors (RMSE), structural similarity index measure (SSIM) and gamma index on 10 specific prostate cancer patients. The significance level was evaluated by t-test with the p-value threshold of 0.05 compared with the results from the group model.

RESULTS

The patient-specific model achieved an average RMSE of 0.014 (p < 0.05 ${{{p}}}<{0.05}$ ), and an average SSIM of 0.981 (p < 0.05 ${{{p}}}<{0.05}$ ), out-performing the group model. Qualitative results also demonstrated that our patient-specific approach acquired better imaging quality with more details reconstructed when comparing with the group model. Dose verification achieved an average RMSE of 0.011 (p < 0.05 ${{{p}}}<{0.05}$ ), and an average SSIM of 0.995 (p < 0.05 ${{{p}}}<{0.05}$ ). Gamma index evaluation demonstrated a high agreement (97.4% [p < 0.05 ${{{p}}}<{0.05}$ ] and 97.9% [p < 0.05 ${{{p}}}<{0.05}$ ] for 1%/3  and 1%/5 mm) between the predicted and the ground truth dose maps. Our approach approximately took 6 s to reconstruct PA images for each patient, demonstrating its feasibility for online 3D dose verification for prostate proton therapy.

CONCLUSIONS

Our method demonstrated the feasibility of achieving 3D high-precision PA-based dose verification using patient-specific deep-learning approaches, which can potentially be used to guide the treatment to mitigate the impact of range uncertainty and improve the precision. Further studies are needed to validate the clinical impact of the technique.

Hybrid-supervised deep learning for domain transfer 3D protoacoustic image reconstruction

Objective. Protoacoustic imaging showed great promise in providing real-time 3D dose verification of proton therapy. However, the limited acquisition angle in protoacoustic imaging induces severe artifacts, which impairs its accuracy for dose verification. In this study, we developed a hybrid-supervised deep learning method for protoacoustic imaging to address the limited view issue.Approach. We proposed a Recon-Enhance two-stage deep learning method. In the Recon-stage, a transformer-based network was developed to reconstruct initial pressure maps from raw acoustic signals. The network is trained in a hybrid-supervised approach, where it is first trained using supervision by the iteratively reconstructed pressure map and then fine-tuned using transfer learning and self-supervision based on the data fidelity constraint. In the enhance-stage, a 3D U-net is applied to further enhance the image quality with supervision from the ground truth pressure map. The final protoacoustic images are then converted to dose for proton verification.Main results. The results evaluated on a dataset of 126 prostate cancer patients achieved an average root mean squared errors (RMSE) of 0.0292, and an average structural similarity index measure (SSIM) of 0.9618, out-performing related start-of-the-art methods. Qualitative results also demonstrated that our approach addressed the limit-view issue with more details reconstructed. Dose verification achieved an average RMSE of 0.018, and an average SSIM of 0.9891. Gamma index evaluation demonstrated a high agreement (94.7% and 95.7% for 1%/3 mm and 1%/5 mm) between the predicted and the ground truth dose maps. Notably, the processing time was reduced to 6 s, demonstrating its feasibility for online 3D dose verification for prostate proton therapy.Significance. Our study achieved start-of-the-art performance in the challenging task of direct reconstruction from radiofrequency signals, demonstrating the great promise of PA imaging as a highly efficient and accurate tool forinvivo3D proton dose verification to minimize the range uncertainties of proton therapy to improve its precision and outcomes.

Machine learning approach for proton range verification using real-time prompt gamma imaging with Compton cameras: addressing the total deposited energy information gap

Objective.Compton camera imaging shows promise as a range verification technique in proton therapy. This work aims to assess the performance of a machine learning model in Compton camera imaging for proton beam range verification improvement.Approach.The presented approach was used to recognize Compton events and estimate more accurately the prompt gamma (PG) energy in the Compton camera to reconstruct the PGs emission profile during proton therapy. This work reports the results obtained from the Geant4 simulation for a proton beam impinging on a polymethyl methacrylate (PMMA) target. To validate the versatility of such an approach, the produced PG emissions interact with a scintillating fiber-based Compton camera.Main results.A trained multilayer perceptron (MLP) neural network shows that it was possible to achieve a notable three-fold increase in the signal-to-total ratio. Furthermore, after event selection by the trained MLP, the loss of full-energy PGs was compensated by means of fitting an MLP energy regression model to the available data from true Compton (signal) events, predicting more precisely the total deposited energy for Compton events with incomplete energy deposition.Significance.A considerable improvement in the Compton camera's performance was demonstrated in determining the distal falloff and identifying a few millimeters of target displacements. This approach has shown great potential for enhancing online proton range monitoring with Compton cameras in future clinical applications.

Gamma-ray sources imaging and test-beam results with MACACO III Compton camera

Hadron therapy is a radiotherapy modality which offers a precise energy deposition to the tumors and a dose reduction to healthy tissue as compared to conventional methods. However, methods for real-time monitoring are required to ensure that the radiation dose is deposited on the target. The IRIS group of IFIC-Valencia developed a Compton camera prototype for this purpose, intending to image the Prompt Gammas emitted by the tissue during irradiation. The system detectors are composed of Lanthanum (III) bromide scintillator crystals coupled to silicon photomultipliers. After an initial characterization in the laboratory, in order to assess the system capabilities for future experiments in proton therapy centers, different tests were carried out in two facilities: PARTREC (Groningen, The Netherlands) and the CNA cyclotron (Sevilla, Spain). Characterization studies performed at PARTREC indicated that the detectors linearity was improved with respect to the previous version and an energy resolution of 5.2 % FWHM at 511 keV was achieved. Moreover, the imaging capabilities of the system were evaluated with a line source of 68Ge and a point-like source of 241Am-9Be. Images at 4.439 MeV were obtained from irradiation of a graphite target with an 18 MeV proton beam at CNA, to perform a study of the system potential to detect shifts at different intensities. In this sense, the system was able to distinguish 1 mm variations in the target position at different beam current intensities for measurement times of 1800 and 600 s.

Range verification in heavy-ion therapy using a hadron tumour marker

Objective.A new method to estimate the range of an ion beam in a patient during heavy-ion therapy was investigated, which was previously verified for application in proton therapy.Approach.The method consists of placing a hadron tumour marker (HTM) close to the tumour. As the treatment beam impinges on the HTM, the marker undergoes nuclear reactions. When the HTM material is carefully chosen, the activation results in the emission of several delayed, characteristicγrays, whose intensities are correlated with the remaining range inside the patient. When not just one but two reaction channels are investigated, the ratio between these twoγray emissions can be measured, and the ratio is independent of any beam delivery uncertainties.Main results.A proof-of-principle experiment with an16O ion beam and Ag foils as HTM was successfully executed. The107Ag(16O,x)112Sb and the107Ag(16O,x)114Sb reaction channels were identified as suitable for the HTM technique. When only oneγ-ray emission is measured, the resulting range-uncertainty estimation is at the 0.5 mm scale. When both channels are considered, a theoretical limit on the range uncertainty of a clinical fiducal marker was found to be ±290μm.Significance.Range uncertainty of a heavy-ion beam limits the prescribed treatment plan for cancer patients, especially the direction of the ion beam in relation to any organ at risk. An easy to implement range-verification technique which can be utilized during clinical treatment would allow treatment plans to take full advantage of the sharp fall-off of the Bragg peak without the risk of depositing excessive dose into healthy tissue.

Uncertainty-aware spot rejection rate as quality metric for proton therapy using a digital tracking calorimeter

Objective.Proton therapy is highly sensitive to range uncertainties due to the nature of the dose deposition of charged particles. To ensure treatment quality, range verification methods can be used to verify that the individual spots in a pencil beam scanning treatment fraction match the treatment plan. This study introduces a novel metric for proton therapy quality control based on uncertainties in range verification of individual spots.Approach.We employ uncertainty-aware deep neural networks to predict the Bragg peak depth in an anthropomorphic phantom based on secondary charged particle detection in a silicon pixel telescope designed for proton computed tomography. The subsequently predicted Bragg peak positions, along with their uncertainties, are compared to the treatment plan, rejecting spots which are predicted to be outside the 95% confidence interval. The such-produced spot rejection rate presents a metric for the quality of the treatment fraction.Main results.The introduced spot rejection rate metric is shown to be well-defined for range predictors with well-calibrated uncertainties. Using this method, treatment errors in the form of lateral shifts can be detected down to 1 mm after around 1400 treated spots with spot intensities of 1 × 107protons. The range verification model used in this metric predicts the Bragg peak depth to a mean absolute error of 1.107 ± 0.015 mm.Significance.Uncertainty-aware machine learning has potential applications in proton therapy quality control. This work presents the foundation for future developments in this area.

Influence of the background in Compton camera images for proton therapy treatment monitoring

Objective. Background events are one of the most relevant contributions to image degradation in Compton camera imaging for hadron therapy treatment monitoring. A study of the background and its contribution to image degradation is important to define future strategies to reduce the background in the system.Approach. In this simulation study, the percentage of different kinds of events and their contribution to the reconstructed image in a two-layer Compton camera have been evaluated. To this end, GATE v8.2 simulations of a proton beam impinging on a PMMA phantom have been carried out, for different proton beam energies and at different beam intensities.Main results. For a simulated Compton camera made of Lanthanum (III) Bromide monolithic crystals, coincidences caused by neutrons arriving from the phantom are the most common type of background produced by secondary radiations in the Compton camera, causing between 13% and 33% of the detected coincidences, depending on the beam energy. Results also show that random coincidences are a significant cause of image degradation at high beam intensities, and their influence in the reconstructed images is studied for values of the time coincidence windows from 500 ps to 100 ns.Significance. Results indicate the timing capabilities required to retrieve the fall-off position with good precision. Still, the noise observed in the image when no randoms are considered make us consider further background rejection methods.

Hybrid PET/Compton-camera imaging: an imager for the next generation

Compton cameras can offer advantages over gamma cameras for some applications, since they are well suited for multitracer imaging and for imaging high-energy radiotracers, such as those employed in radionuclide therapy. While in conventional clinical settings state-of-the-art Compton cameras cannot compete with well-established methods such as PET and SPECT, there are specific scenarios in which they can constitute an advantageous alternative. The combination of PET and Compton imaging can benefit from the improved resolution and sensitivity of current PET technology and, at the same time, overcome PET limitations in the use of multiple radiotracers. Such a system can provide simultaneous assessment of different radiotracers under identical conditions and reduce errors associated with physical factors that can change between acquisitions. Advances are being made both in instrumentation developments combining PET and Compton cameras for multimodal or three-gamma imaging systems, and in image reconstruction, addressing the challenges imposed by the combination of the two modalities or the new techniques. This review article summarizes the advances made in Compton cameras for medical imaging and their combination with PET.

Prompt-gamma fall-off estimation with C-ion irradiation at clinical energies, using a knife-edge slit camera: A Monte Carlo study

PURPOSE

In-vivo range verification has been a hot topic in particle therapy since two decades. Many efforts have been done for proton therapy, while fewer studies were conducted considering a beam of carbon ions. In the present work, a simulation study was performed to show whether it is possible to measure the prompt-gamma fall-off inside the high neutron background typical of carbon-ion irradiation, using a knife-edge slit camera. In addition to this, we wanted to estimate the uncertainty in retrieving the particle range in the case of a pencil beam of C-ions at clinically relevant energy of 150 MeVu.

METHODS

For these purposes, the Monte Carlo code FLUKA was adopted for simulations and three different analytical methods were implemented to get the accuracy in the range retrieval of the simulated set-up.

RESULTS

The analysis of simulation data has brought to the promising and desired precision of about 4 mm in the determination of the dose profile fall-off in case of a spill irradiation, for which all the three cited methods were coherent in their predictions.

CONCLUSIONS

The Prompt Gamma Imaging technique should be further studied as a tool to reduce range uncertainties affecting carbon ion radiation therapy.

Development of a two-layer dense-pixel LYSO Compton camera prototype for prompt gamma imaging

Objective.Lutetium-yttrium orthosilicate (LYSO)-based Compton camera (CC) has been proposed for prompt gamma imaging due to its high detection efficiency and position resolution. However, very few LYSO CC prototypes have been built and used for practical evaluation. In this study, we built a lightweight dense-pixel silicon photomultiplier-based two-layer LYSO CC prototype for future prompt gamma imaging.Approach.We attempt the first-ever effort to use the double-encoding with the thick light guide and coding circuit structure for 46 × 46 dense-pixel LYSO detectors construction and use pixel segmentation based on centroid mapping to obtain 4232 spectral calibrations. We also present a framework for list-mode projection data acquisition based on the decoding of the time series data obtained by data acquisition card in this study. Finally, the standard source calibration, ring-like²²Na source with non-uniform intensity, and mixed point-like source with a wide energy spectrum experiments were implemented to evaluate the resolution metrics and imaging performance of the prototype.Main results.The lateral position resolution of the prototype was 1 mm, and the maximum measurement deviation is 2.5 mm and 5 mm in the depth direction for the scatterer and absorber, respectively. In the experiments, the measured energy resolution was 9.63% @ 1.33 MeV for the scatterer and 10.8% @ 1.33 MeV for the absorber. And the detection efficiency of the prototype for a spherical⁶⁰Co source with a diameter of 2.8 mm at 10 cm far was 5.7 × 10^-3@ 1.33 MeV and the full width at half maximum of the reconstruction was 5.5 mm. Besides, the spatial position offset within 2 mm of the radioactive source at 10 cm can be distinguished.Signification.The developed two-layer dense-pixel LYSO CC contributes to incorporating Compton imaging techniques for prompt gamma detection and multiple energy sources into nuclear medical imaging.

Design and performance evaluation of a slit-slat camera for 2D prompt gamma imaging in proton therapy monitoring: A Monte Carlo simulation study

PURPOSE

We investigated the design of a prompt gamma camera for real-time dose delivery verification and the partial mitigation of range uncertainties.

METHODS

A slit slat (SS) camera was optimized using the trade-off between the signal-to-noise ratio and spatial resolution. Then, using the GATE Monte Carlo package, the camera performances were estimated by means of target shifts, beam position quantification, changing the camera distance from the beam, and air cavity inserting. A homogeneous PMMA phantom and the air gaps induced PMMA phantom were used. The air gaps ranged from 5 mm to 30 mm by 5 mm increments were positioned in the middle of the beam range. To reduce the simulation time, phase space scoring was used. The batch method with five realizations was used for stochastic error calculations.

RESULTS

The system's detection efficiency was 1.1 × 10 - 4 PGs Emitted PGs ( 1.8 × 10 - 5 $1.1 \times {10}^{-4}\frac{{\rm PGs}}{{\rm Emitted}\ {\rm PGs}}\ (1.8 \times {10}^{-5}$ PGs/proton) for a 10 × 20 cm² detector (source-to-collimator distance = 15.0 cm). Axial and transaxial resolutions were 23 mm and 18 mm, respectively. The SS camera estimated the range as 69.0 ± 3.4 (relative stochastic error 1-sigma is 5%) and 67.6 ± 1.8 mm (2.6%) for the real range of 67.0 mm for 10⁷ and 10⁸ protons of 100 MeV, respectively. Considering 160 MeV, these values are 155.5 ± 3.1 (2%) and 152.2 ± 2.0 mm (1.3%) for the real range of 152.0 mm for 10⁷ and 10⁸ protons, respectively. Considering phantom shift, for a 100 MeV beam, the precision of the quantification (1-sigma) in the axial and lateral phantom shift estimation is 2.6 mm and 1 mm, respectively. Accordingly, the axial and lateral quantification precisions were 1.3 mm and 1 mm for a 160 MeV beam, respectively. Furthermore, the quantification of an air gap formulated as gap d e t = 0.98 × gap real ${{\rm gap}}_{det}=0.98 \times {{\rm gap}}_{{\rm real}}$ , where gap d e t ${{\rm gap}}_{det}$ and gap_real are the estimated and real air gap, respectively. The precision of the air gap quantification is 1.6 mm (1 sigma). Moreover, 2D PG images show the trajectory of the proton beam through the phantom.

CONCLUSION

The proposed slit-slat imaging systems can potentially provide a real-time, in-vivo, and non-invasive treatment monitoring method for proton therapy.

3Din vivodose verification in prostate proton therapy with deep learning-based proton-acoustic imaging

Dose delivery uncertainty is a major concern in proton therapy, adversely affecting the treatment precision and outcome. Recently, a promising technique, proton-acoustic (PA) imaging, has been developed to provide real-timein vivo3D dose verification. However, its dosimetry accuracy is limited due to the limited-angle view of the ultrasound transducer. In this study, we developed a deep learning-based method to address the limited-view issue in the PA reconstruction. A deep cascaded convolutional neural network (DC-CNN) was proposed to reconstruct 3D high-quality radiation-induced pressures using PA signals detected by a matrix array, and then derive precise 3D dosimetry from pressures for dose verification in proton therapy. To validate its performance, we collected 81 prostate cancer patients' proton therapy treatment plans. Dose was calculated using the commercial software RayStation and was normalized to the maximum dose. The PA simulation was performed using the open-source k-wave package. A matrix ultrasound array with 64 × 64 sensors and 500 kHz central frequency was simulated near the perineum to acquire radiofrequency (RF) signals during dose delivery. For realistic acoustic simulations, tissue heterogeneity and attenuation were considered, and Gaussian white noise was added to the acquired RF signals. The proposed DC-CNN was trained on 204 samples from 69 patients and tested on 26 samples from 12 other patients. Predicted 3D pressures and dose maps were compared against the ground truth qualitatively and quantitatively using root-mean-squared-error (RMSE), gamma-index (GI), and dice coefficient of isodose lines. Results demonstrated that the proposed method considerably improved the limited-view PA image quality, reconstructing pressures with clear and accurate structures and deriving doses with a high agreement with the ground truth. Quantitatively, the pressure accuracy achieved an RMSE of 0.061, and the dose accuracy achieved an RMSE of 0.044, GI (3%/3 mm) of 93.71%, and 90%-isodose line dice of 0.922. The proposed method demonstrates the feasibility of achieving high-quality quantitative 3D dosimetry in PA imaging using a matrix array, which potentially enables the online 3D dose verification for prostate proton therapy.

Development and Applications of Compton Camera-A Review

The history of Compton cameras began with the detection of radiation sources originally for applications in astronomy. A Compton camera is a promising γ-ray detector that operates in the wide energy range of a few tens of keV to MeV. The γ-ray detection method of a Compton camera is based on Compton scattering kinematics, which is used to determine the direction and energy of the γ-rays without using a mechanical collimator. Although the Compton camera was originally designed for astrophysical applications, it was later applied in medical imaging as well. Moreover, its application in environmental radiation measurements is also under study. Although a few review papers regarding Compton cameras have been published, they either focus very specifically on the detectors used in such cameras or the particular applications of Compton cameras. Thus, the aim of this paper is to review the features and types of Compton cameras and introduce their applications, associated imaging algorithms, improvement scopes, and their future aspects.

Feasibility of triple gamma ray imaging of 10C for range verification in ion therapy

Objective. In carbon ion therapy, the visualization of the range of incident particles in a patient body is important for treatment verification. In-beam positron emission tomography (PET) imaging is one of the methods to verify the treatment in ion therapy due to the high quality of PET images. We have shown the feasibility of in-beam PET imaging of radioactive 15O and 11C ion beams for range verification using our OpenPET system. Recently, we developed a whole gamma imager (WGI) that can simultaneously work as PET, single gamma ray and triple gamma ray imaging. The WGI has high potential to detect the location of 10C, which emits positrons with a simultaneous gamma ray of 718 keV, within the patient’s body during ion therapy. Approach. In this work, we focus on investigating the performance of WGI for 10C imaging and its feasibility for range verification in carbon ion therapy. First, the performance of the WGI was studied to image a 10C point source using the Geant4 toolkit. Then, the feasibility of WGI was investigated for an irradiated polymethyl methacrylate (PMMA) phantom with a 10C ion beam at the carbon therapy facility of the Heavy Ion Medical Accelerator in Chiba. Main results. The average spatial resolution and sensitivity for the simulated 10C point source at the centre of the field of view were 5.5 mm FWHM and 0.010%, respectively. The depth dose of the 10C ion beam was measured, and the triple gamma image of 10C nuclides for an irradiated PMMA phantom was obtained by applying a simple back projection to the detected triple gammas. Significance. The shift between Bragg peak position and position of the peak of the triple gamma image in an irradiated PMMA phantom was 2.8 ± 0.8 mm, which demonstrates the capability of triple gamma imaging using WGI for range verification of 10C ion beams.

Joint image reconstruction algorithm in Compton cameras

Objective. To demonstrate the benefits of using an joint image reconstruction algorithm based on the List Mode Maximum Likelihood Expectation Maximization that combines events measured in different channels of information of a Compton camera. Approach. Both simulations and experimental data are employed to show the algorithm performance. Main results. The obtained joint images present improved image quality and yield better estimates of displacements of high-energy gamma-ray emitting sources. The algorithm also provides images that are more stable than any individual channel against the noisy convergence that characterizes Maximum Likelihood based algorithms. Significance. The joint reconstruction algorithm can improve the quality and robustness of Compton camera images. It also has high versatility, as it can be easily adapted to any Compton camera geometry. It is thus expected to represent an important step in the optimization of Compton camera imaging.

Machine learning-based event recognition in SiFi Compton camera imaging for proton therapy monitoring

Objective. Online monitoring of dose distribution in proton therapy is currently being investigated with the detection of prompt gamma (PG) radiation emitted from a patient during irradiation. The SiPM and scintillation Fiber based Compton Camera (SiFi-CC) setup is being developed for this aim. Approach. A machine learning approach to recognize Compton events is proposed, reconstructing the PG emission profile during proton therapy. The proposed method was verified on pseudo-data generated by a Geant4 simulation for a single proton beam impinging on a polymethyl methacrylate (PMMA) phantom. Three different models including the boosted decision tree (BDT), multilayer perception (MLP) neural network, and k-nearest neighbour (k-NN) were trained using 10-fold cross-validation and then their performances were assessed using the receiver operating characteristic (ROI) curves. Subsequently, after event selection by the most robust model, a software based on the List-Mode Maximum Likelihood Estimation Maximization (LM-MLEM) algorithm was applied for the reconstruction of the PG emission distribution profile. Main results. It was demonstrated that the BDT model excels in signal/background separation compared to the other two. Furthermore, the reconstructed PG vertex distribution after event selection showed a significant improvement in distal falloff position determination. Significance. A highly satisfactory agreement between the reconstructed distal edge position and that of the simulated Compton events was achieved. It was also shown that a position resolution of 3.5 mm full width at half maximum (FWHM) in distal edge position determination is feasible with the proposed setup.

Compton imaging for medical applications

Compton imaging exploits inelastic scattering, known as Compton scattering, using a Compton camera consisting of a scatterer detector in the front layer and an absorber detector in the back layer. This method was developed for astronomy, and in recent years, research and development for environmental and medical applications has been actively conducted. Compton imaging can discriminate gamma rays over a wide energy range from several hundred keV to several MeV. Therefore, it is expected to be applied to the simultaneous imaging of multiple nuclides in nuclear medicine and prompt gamma ray imaging for range verification in particle therapy. In addition, multiple gamma coincidence imaging is expected to be realized, which allows the source position to be determined from a single coincidence event using nuclides that emit multiple gamma rays simultaneously, such as nuclides that emit a single gamma ray simultaneously with positron decay. This review introduces various efforts toward the practical application of Compton imaging in the medical field, including in vivo studies, and discusses its prospects.

Range uncertainty reductions in proton therapy may lead to the feasibility of novel beam arrangements which improve organ-at-risk sparing

PURPOSE

In proton therapy, dose distributions are currently often conformed to organs at risk (OARs) using the less sharp dose fall-off at the lateral beam edge to reduce the effects of uncertainties in the in vivo proton range. However, range uncertainty reductions may make greater use of the sharper dose fall-off at the distal beam edge feasible, potentially improving OAR sparing. We quantified the benefits of such novel beam arrangements.

METHODS

For each of 10 brain or skull base cases, five treatment plans robust to 2 mm setup and 0%-4% range uncertainty were created for the traditional clinical beam arrangement and a novel beam arrangement making greater use of the distal beam edge to conform the dose distribution to the brainstem. Metrics including the brainstem normal tissue complication probability (NTCP) with the endpoint of necrosis were determined for all plans and all setup and range uncertainty scenarios.

RESULTS

For the traditional beam arrangement, reducing the range uncertainty from the current level of approximately 4% to a potentially achievable level of 1% reduced the brainstem NTCP by up to 0.9 percentage points in the nominal and up to 1.5 percentage points in the worst-case scenario. Switching to the novel beam arrangement at 1% range uncertainty improved these values by a factor of 2, that is, to 1.8 percentage points and 3.2 percentage points, respectively. The novel beam arrangement achieved a lower brainstem NTCP in all cases starting at a range uncertainty of 2%.

CONCLUSION

The benefits of novel beam arrangements may be of the same magnitude or even exceed the direct benefits of range uncertainty reductions. Indirect effects may therefore contribute markedly to the benefits of reducing proton range uncertainties.

Applications of Machine Learning to Improve the Clinical Viability of Compton Camera Based in vivo Range Verification in Proton Radiotherapy

We studied the application of a deep, fully connected Neural Network (NN) to process prompt gamma (PG) data measured by a Compton camera (CC) during the delivery of clinical proton radiotherapy beams. The network identifies 1) recorded "bad" PG events arising from background noise during the measurement, and 2) the correct ordering of PG interactions in the CC to help improve the fidelity of "good" data used for image reconstruction. PG emission from a tissue-equivalent target during irradiation with a 150 MeV proton beam delivered at clinical dose rates was measured with a prototype CC. Images were reconstructed from both the raw measured data and the measured data that was further processed with a neural network (NN) trained to identify "good" and "bad" PG events and predict the ordering of individual interactions within the good PG events. We determine if NN processing of the CC data could improve the reconstructed PG images to a level in which they could provide clinically useful information about the in vivo range and range shifts of the proton beams delivered at full clinical dose rates. Results showed that a deep, fully connected NN improved the achievable contrast to noise ratio (CNR) in our images by more than a factor of 8x. This allowed the path, range, and lateral width of the clinical proton beam within a tissue equivalent target to easily be identified from the PG images, even at the highest dose rates of a 150 MeV proton beam used for clinical treatments. On average, shifts in the beam range as small as 3 mm could be identified. However, when limited by the amount of PG data measured with our prototype CC during the delivery of a single proton pencil beam (~1 × 109 protons), the uncertainty in the reconstructed PG images limited the identification of range shift to ~5 mm. Substantial improvements in CC images were obtained during clinical beam delivery through NN pre-processing of the measured PG data. We believe this shows the potential of NNs to help improve and push CC-based PG imaging toward eventual clinical application for proton RT treatment delivery verification.

Radiopharmaceutical imaging based on 3D-CZT Compton camera with 3D-printed mouse phantom

PURPOSE

The study proposes the use of three-dimensional CdZnTe Compton camera (3D-CZT CC) for radiopharmaceutical imaging and investigates the influence factors using a 3D-printed mouse phantom.

METHODS

The event selection method and image reconstruction algorithm are optimized by Monte Carlo simulations and mouse phantom experiments.

RESULTS

Simulation results show that the intrinsic energy resolution and spatial resolution of 3D-CZT cause a certain deviation in the calculated Compton scattering angle and Compton axis. Such deviation causes the imaging quality to deteriorate. By selecting events whose distance between Compton and photoelectronic interactions are larger than 10 mm, the mean deviation of the Compton axis could be reduced to less than 10%. Using the ordered origin ensemble algorithm with resolution recovery, the artifacts around organs where the radiopharmaceutical was placed are reduced, and the quality of the reconstruction results are improved compared to the results with simple back projection and origin ensembles algorithms. The phantom study shows that the 3D-CZT CC imaging device could visualize the radiopharmaceuticals distribution by 15 min detection.

CONCLUSIONS

Through the analysis of this study, the feasibility of 3D-CZT CC for in-vivo distribution measurement of radiopharmaceuticals is demonstrated, and the quality of reconstruction result has been improved.

The effects of Compton camera data acquisition and readout timing on PG imaging for proton range verification

The purpose of this study was to determine how the characteristics of the data acquisition (DAQ) electronics of a Compton camera (CC) affect the quality of the recorded prompt gamma (PG) interaction data and the reconstructed images, during clinical proton beam delivery. We used the Monte-Carlo-plus-Detector-Effect (MCDE) model to simulate the delivery of a 150 MeV clinical proton pencil beam to a tissue-equivalent plastic phantom. With the MCDE model we analyzed how the recorded PG interaction data changed as two characteristics of the DAQ electronics of a CC were changed: (1) the number of data readout channels; and (2) the active charge collection, readout, and reset time. As the proton beam dose rate increased, the number of recorded PG single-, double-, and triple-scatter events decreased by a factor of 60× for the current DAQ configuration of the CC. However, as the DAQ readout channels were increased and the readout/reset timing decreased, the number of recorded events decreased by <5× at the highest clinical dose rate. The increased number of readout channels and reduced readout/reset timing also resulted in higher quality recorded data. That is, a higher percentage of the recorded double- and triple-scatters were "true" events (caused by a single incident gamma) and not "false" events (caused by multiple incident gammas). The increase in the number and the quality of recorded data allowed higher quality PG images to be reconstructed even at the highest clinical dose rates.

Towards machine learning aided real-time range imaging in proton therapy

Compton imaging represents a promising technique for range verification in proton therapy treatments. In this work, we report on the advantageous aspects of the i-TED detector for proton-range monitoring, based on the results of the first Monte Carlo study of its applicability to this field. i-TED is an array of Compton cameras, that have been specifically designed for neutron-capture nuclear physics experiments, which are characterized by [Formula: see text]-ray energies spanning up to 5-6 MeV, rather low [Formula: see text]-ray emission yields and very intense neutron induced [Formula: see text]-ray backgrounds. Our developments to cope with these three aspects are concomitant with those required in the field of hadron therapy, especially in terms of high efficiency for real-time monitoring, low sensitivity to neutron backgrounds and reliable performance at the high [Formula: see text]-ray energies. We find that signal-to-background ratios can be appreciably improved with i-TED thanks to its light-weight design and the low neutron-capture cross sections of its LaCl[Formula: see text] crystals, when compared to other similar systems based on LYSO, CdZnTe or LaBr[Formula: see text]. Its high time-resolution (CRT [Formula: see text] 500 ps) represents an additional advantage for background suppression when operated in pulsed HT mode. Each i-TED Compton module features two detection planes of very large LaCl[Formula: see text] monolithic crystals, thereby achieving a high efficiency in coincidence of 0.2% for a point-like 1 MeV [Formula: see text]-ray source at 5 cm distance. This leads to sufficient statistics for reliable image reconstruction with an array of four i-TED detectors assuming clinical intensities of 10[Formula: see text] protons per treatment point. The use of a two-plane design instead of three-planes has been preferred owing to the higher attainable efficiency for double time-coincidences than for threefold events. The loss of full-energy events for high energy [Formula: see text]-rays is compensated by means of machine-learning based algorithms, which allow one to enhance the signal-to-total ratio up to a factor of 2.

Proton range monitoring using 13N peak for proton therapy applications

The Monte Carlo method is employed in this study to simulate the proton irradiation of a water-gel phantom. Positron-emitting radionuclides such as ¹¹C, ¹⁵O, and ¹³N are scored using the Particle and Heavy Ion Transport Code System Monte Carlo code package. Previously, it was reported that as a result of ¹⁶O(p,2p2n)¹³N nuclear reaction, whose threshold energy is relatively low (5.660 MeV), a ¹³N peak is formed near the actual Bragg peak. Considering the generated ¹³N peak, we obtain offset distance values between the ¹³N peak and the actual Bragg peak for various incident proton energies ranging from 45 to 250 MeV, with an energy interval of 5 MeV. The offset distances fluctuate between 1.0 and 2.0 mm. For example, the offset distances between the ¹³N peak and the Bragg peak are 2.0, 2.0, and 1.0 mm for incident proton energies of 80, 160, and 240 MeV, respectively. These slight fluctuations for different incident proton energies are due to the relatively stable energy-dependent cross-section data for the ¹⁶O(p,2p2n)¹³N nuclear reaction. Hence, we develop an open-source computer program that performs linear and non-linear interpolations of offset distance data against the incident proton energy, which further reduces the energy interval from 5 to 0.1 MeV. In addition, we perform spectral analysis to reconstruct the ¹³N Bragg peak, and the results are consistent with those predicted from Monte Carlo computations. Hence, the results are used to generate three-dimensional scatter plots of the ¹³N radionuclide distribution in the modeled phantom. The obtained results and the developed methodologies will facilitate future investigations into proton range monitoring for therapeutic applications.

Dual-layer spectral CT for proton, helium, and carbon ion beam therapy planning of brain tumors

Pretreatment computed tomography (CT) imaging is an essential component of the particle therapy treatment planning chain. Treatment planning and optimization with charged particles require accurate and precise estimations of ion beam range in tissues, characterized by the stopping power ratio (SPR). Reduction of range uncertainties arising from conventional CT-number-to-SPR conversion based on single-energy CT (SECT) imaging is of importance for improving clinical practice. Here, the application of a novel imaging and computational methodology using dual-layer spectral CT (DLCT) was performed toward refining patient-specific SPR estimates. A workflow for DLCT-based treatment planning was devised to evaluate SPR prediction for proton, helium, and carbon ion beam therapy planning in the brain. DLCT- and SECT-based SPR predictions were compared in homogeneous and heterogeneous anatomical regions. This study included eight patients scanned for diagnostic purposes with a DLCT scanner. For each patient, four different treatment plans were created, simulating tumors in different parts of the brain. For homogeneous anatomical regions, mean SPR differences of about 1% between the DLCT- and SECT-based approaches were found. In plans of heterogeneous anatomies, relative (absolute) proton range shifts of 0.6% (0.4 mm) in the mean and up to 4.4% (2.1 mm) at the distal fall-off were observed. In the investigated cohort, 12% of the evaluated organs-at-risk (OARs) presented differences in mean or maximum dose of more than 0.5 Gy (RBE) and up to 6.8 Gy (RBE) over the entire treatment. Range shifts and dose differences in OARs between DLCT and SECT in helium and carbon ion treatment plans were similar to protons. In the majority of investigated cases (75th percentile), SECT- and DLCT-based range estimations were within 0.6 mm. Nonetheless, the magnitude of patient-specific range deviations between SECT and DLCT was clinically relevant in heterogeneous anatomical sites, suggesting further study in larger, more diverse cohorts. Results indicate that patients with brain tumors may benefit from DLCT-based treatment planning.

Carbon range verification with 718 keV Compton imaging

Carbon ion radiotherapy is a sophisticated radiation treatment modality because of its superiority in achieving precise dosage distribution and high biological effectiveness. However, there exist beam range uncertainties that affect treatment efficiency. This problem can be resolved if the clinical beam could be monitored precisely in real-time, such as by imaging the prompt gamma emission from the target. In this study, we performed real-time detection and imaging of 718 keV prompt gamma emissions using a Si/CdTe Compton camera. We conducted experiments on graphite phantoms using clinical carbon ion beams of 290 MeV/u energy. Compton images were reconstructed using simple back-projection methods from the energy events of 718 keV prompt gamma emissions. The peak intensity position in reconstructed 718 keV prompt gamma images was few millimeters below the Bragg peak position. Moreover, the dual- and triple-energy window images for all positions of phantoms were not affected by scattered gammas, and their peak intensity positions were approximately similar to those observed in the reconstructed 718 keV prompt gamma images. In conclusion, the findings of the current study demonstrate the feasibility of using our Compton camera for real-time beam monitoring of carbon ion beams under clinical beam intensity.

Proton range uncertainty reduction benefits for skull base tumors in terms of normal tissue complication probability (NTCP) and healthy tissue doses

PURPOSE

Proton therapy allows for more conformal dose distributions and lower organ at risk and healthy tissue doses than conventional photon-based radiotherapy, but uncertainties in the proton range currently prevent proton therapy from making full use of these advantages. Numerous developments therefore aim to reduce such range uncertainties. In this work, we quantify the benefits of reductions in range uncertainty for treatments of skull base tumors.

METHODS

The study encompassed 10 skull base patients with clival tumors. For every patient, six treatment plans robust to setup errors of 2 mm and range errors from 0% to 5% were created. The determined metrics included the brainstem and optic chiasm normal tissue complication probability (NTCP) with the endpoints of necrosis and blindness, respectively, as well as the healthy tissue volume receiving at least 70% of the prescription dose.

RESULTS

A range uncertainty reduction from the current level of 4% to a potentially achievable level of 1% reduced the probability of brainstem necrosis by up to 1.3 percentage points in the nominal scenario in which neither setup nor range errors occur and by up to 2.9 percentage points in the worst-case scenario. Such a range uncertainty reduction also reduced the optic chiasm NTCP with the endpoint of blindness by up to 0.9 percentage points in the nominal scenario and by up to 2.2 percentage points in the worst-case scenario. The decrease in the healthy tissue volume receiving at least 70% of the prescription dose ranged from -7.8 to 24.1 cc in the nominal scenario and from -3.4 to 38.4 cc in the worst-case scenario.

CONCLUSION

The benefits quantified as part of this study serve as a guideline of the OAR and healthy tissue dose benefits that range monitoring techniques may be able to achieve. Benefits were observed between all levels of range uncertainty. Even smaller range uncertainty reductions may therefore be beneficial.

Towards real-time PGS range monitoring in proton therapy of prostate cancer

Proton therapy of prostate cancer (PCPT) was linked with increased levels of gastrointestinal toxicity in its early use compared to intensity-modulated radiation therapy (IMRT). The higher radiation dose to the rectum by proton beams is mainly due to anatomical variations. Here, we demonstrate an approach to monitor rectal radiation exposure in PCPT based on prompt gamma spectroscopy (PGS). Endorectal balloons (ERBs) are used to stabilize prostate movement during radiotherapy. These ERBs are usually filled with water. However, other water solutions containing elements with higher atomic numbers, such as silicon, may enable the use of PGS to monitor the radiation exposure of the rectum. Protons hitting silicon atoms emit prompt gamma rays with a specific energy of 1.78 MeV, which can be used to monitor whether the ERB is being hit. In a binary approach, we search the silicon energy peaks for every irradiated prostate region. We demonstrate this technique for both single-spot irradiation and real treatment plans. Real-time feedback based on the ERB being hit column-wise is feasible and would allow clinicians to decide whether to adapt or continue treatment. This technique may be extended to other cancer types and organs at risk, such as the oesophagus.

Prompt-gamma emission in GEANT4 revisited and confronted with experiment

PURPOSE

The field of online monitoring of the beam range is one of the most researched topics in proton therapy over the last decade. The development of detectors that can be used for beam range verification under clinical conditions is a challenging task. One promising possible solution are modalities that record prompt-gamma radiation produced by the interactions of the proton beam with the target tissue. A good understanding of the energy spectra of the prompt gammas and the yields in certain energy regions is crucial for a successful design of a prompt-gamma detector. Monte-Carlo simulations are an important tool in development and testing of detector concepts, thus the proper modelling of the prompt-gamma emission in those simulations are of vital importance. In this paper, we confront a number of GEANT4 simulations of prompt-gamma emission, performed with different versions of the package and different physics lists, with experimental data obtained from a phantom irradiation with proton beams of four different energies in the range 70-230 MeV.

METHODS

The comparison is made on different levels: features of the prompt-gamma energy spectrum, gamma emission depth profiles for discrete transitions and the width of the distal fall-off in those profiles.

RESULTS

The best agreement between the measurements and the simulations is found for the GEANT4 version 10.4.2 and the reference physics list QGSP_BIC_HP.

CONCLUSIONS

Modifications to prompt-gamma emission modelling in higher versions of the software increase the discrepancy between the simulation results and the experimental data.

A time-of-flight-based reconstruction for real-time prompt-gamma imaging in proton therapy

We propose a novel prompt-gamma (PG) imaging modality for real-time monitoring in proton therapy: PG time imaging (PGTI). By measuring the time-of-flight (TOF) between a beam monitor and a PG detector, our goal is to reconstruct the PG vertex distribution in 3D. In this paper, a dedicated, non-iterative reconstruction strategy is proposed (PGTI reconstruction). Here, it was resolved under a 1D approximation to measure a proton range shift along the beam direction. In order to show the potential of PGTI in the transverse plane, a second method, based on the calculation of the centre of gravity (COG) of the TIARA pixel detectors' counts was also explored. The feasibility of PGTI was evaluated in two different scenarios. Under the assumption of a 100 ps (rms) time resolution (achievable in single proton regime), MC simulations showed that a millimetric proton range shift is detectable at 2σwith 10⁸incident protons in simplified simulation settings. With the same proton statistics, a potential 2 mm sensitivity (at 2σwith 10⁸incident protons) to beam displacements in the transverse plane was found using the COG method. This level of precision would allow to act in real-time if the treatment does not conform to the treatment plan. A worst case scenario of a 1 ns (rms) TOF resolution was also considered to demonstrate that a degraded timing information can be compensated by increasing the acquisition statistics: in this case, a 2 mm range shift would be detectable at 2σwith 10⁹incident protons. By showing the feasibility of a time-based algorithm for the reconstruction of the PG vertex distribution for a simplified anatomy, this work poses a theoretical basis for the future development of a PG imaging detector based on the measurement of particle TOF.

Influence of sub-nanosecond time of flight resolution for online range verification in proton therapy using the line-cone reconstruction in Compton imaging

Online ion range monitoring in hadron therapy can be performed via detection of secondary radiation, such as promptγ-rays, emitted during treatment. The promptγemission profile is correlated with the ion depth-dose profile and can be reconstructed via Compton imaging. The line-cone reconstruction, using the intersection between the primary beam trajectory and the cone reconstructed via a Compton camera, requires negligible computation time compared to iterative algorithms. A recent report hypothesised that time of flight (TOF) based discrimination could improve the precision of theγfall-off position (FOP) measured via line-cone reconstruction, where TOF comprises both the proton transit time from the phantom entrance untilγemission, and the flight time of theγ-ray to the detector. The aim of this study was to implement such a method and investigate the influence of temporal resolution on the precision of the FOP. Monte Carlo simulations of a 160 MeV proton beam incident on a homogeneous PMMA phantom were performed using GATE. The Compton camera consisted of a silicon-based scatterer and CeBr₃scintillator absorber. The temporal resolution of the detection system (absorber + beam trigger) was varied between 0.1 and 1.3 ns rms and a TOF-based discrimination method applied to eliminate unlikely solution(s) from the line-cone reconstruction. The FOP was obtained for varying temporal resolutions and its precision obtained from its shift across 100 independentγemission profiles compared to a high statistics reference profile. The optimal temporal resolution for the given camera geometry and 10⁸primary protons was 0.2 ns where a precision of 2.30 ± 0.15 mm (1σ) on the FOP was found. This precision is comparable to current state-of-the-art Compton imaging using iterative reconstruction methods or 1D imaging with mechanically collimated devices, and satisfies the requirement of being smaller than the clinical safety margins.

Proton range verification with MACACO II Compton camera enhanced by a neural network for event selection

The applicability extent of hadron therapy for tumor treatment is currently limited by the lack of reliable online monitoring techniques. An active topic of investigation is the research of monitoring systems based on the detection of secondary radiation produced during treatment. MACACO, a multi-layer Compton camera based on LaBr₃ scintillator crystals and SiPMs, is being developed at IFIC-Valencia for this purpose. This work reports the results obtained from measurements of a 150 MeV proton beam impinging on a PMMA target. A neural network trained on Monte Carlo simulations is used for event selection, increasing the signal to background ratio before image reconstruction. Images of the measured prompt gamma distributions are reconstructed by means of a spectral reconstruction code, through which the 4.439 MeV spectral line is resolved. Images of the emission distribution at this energy are reconstructed, allowing calculation of the distal fall-off and identification of target displacements of 3 mm.

Accounting for prompt gamma emission and detection for range verification in proton therapy treatment planning

Prompt gamma (PG) imaging is widely investigated as one of the most promising methods for proton range verification in proton therapy. The performance of this technique is affected by several factors like tissue heterogeneity, number of protons in the considered pencil beam and the detection device. Our previous work proposed a new treatment planning concept which boosts the number of protons of a few PG monitoring-friendly pencil beams (PBs), selected on the basis of two proposed indicators quantifying the conformity between the dose and PG at the emission level, above the desired detectability threshold. To further explore this method at the detection level, in this work we investigated the response of a knife-edge slit PG camera which was deployed in the first clinical application of PG to proton therapy monitoring. The REGistration Graphical User Interface (REGGUI) is employed to simulate the PG emission, PG detection as well as the corresponding dose distribution. As the PG signal detected by this kind of PG camera is sensitive to the relative position of the camera and PG signal falloff, we optimized our PB selection method for this camera by introducing a new camera position indicator identifying whether the expected falloff of the PG signal is centered in the field of view of the camera or not. Our camera-adapted PB selection method is investigated using computed tomography (CT) scans at two different treatment time points of a head and neck, and a prostate cancer patient under scenarios considering different statistics level. The results show that a precision of 0.8 mm for PG falloff identification can be achieved when a PB has more than 2 × 10⁸ primary protons. Except for one case due to unpredictable and comparably large anatomical changes, the PG signals of most of the PBs recommended by all our indicators are observed to be reliable for proton range verification with deviations between the inter-fractional shift of proton range (as deduced from the PB dose distribution) and the detected PG signal within 2.0 mm. In contrast, a shift difference up to 9.6 mm has been observed for the rejected PBs. The magnitude of the proton range shift due to the inter-fractional anatomical changes is observed to be up to 23 mm. The proposed indicators are shown to be valuable for identifying and recommending reliable PBs to create new PG monitoring-friendly TPs. Comparison between our PB boosting method and the alternative PB aggregation, which combines the signal of nearby PBs to reach the desired counting statistics, is also discussed.

Measuring prompt gamma-ray emissions from elements found in tissue during passive-beam proton therapy

Prompt gamma detection during proton radiotherapy for range verification purposes will need to operate in both active and passive treatment beam environments. This paper describes prompt gamma measurements using a high resolution 2″ × 2″ LaBr₃detector for a 200 MeV clinical passive-scatter proton beam. These measurements examine the most likely discrete prompt gamma rays emitted from tissue by detecting gammas produced in water, Perspex, carbon and liquid-nitrogen targets. Measurements were carried out at several positions around the depth corresponding to the location of the Bragg peak for water and Perspex targets in order to investigate prompt gamma emission as a function of depth along the beam path. This work also focused on validating the Geant4 Monte Carlo model of the passive-scatter proton beam line and LaBr₃detector by making a direct comparison between the simulated and experimental results. The initial prompt gamma measurements were overwhelmed by the high amount of scattered radiation when measuring at isocenter, shifting the target further downstream from the final collimator significantly reduced the background radiation. Prompt gamma peaks were then clearly identified for the water, Perspex and graphite targets. The developed Geant4 Monte Carlo model was able to replicate the measured prompt gamma ray energy spectra, including production for important photopeaks to within 10%, except for the 4.44 MeV peak from the water target, which had more than a 50% overestimation of the number of produced prompt gamma rays. The prompt gamma measurements at various depths correlated well with the proton dose deposition; the 4.44 and 6.13 MeV photopeak profiles peaked within 1 cm of the Bragg peak and the R_50%value for the 3-7 MeV energy range predicted the proton range within 8 mm.

Proton therapy range verification method via delayed γ-ray spectroscopy of a molybdenum tumour marker

In this work, a new method of range verification for proton therapy (PT) is experimentally demonstrated for the first time. If a metal marker is implanted near the tumour site, its response to proton activation will result in the emission of characteristic γ rays. The relative intensity of γ rays originating from competing fusion-evaporation reaction channels provides a unique signature of the average proton energy at the marker, and by extension the beam's range, in vivo and in real time. The clinical feasibility of this method was investigated at the PT facility at TRIUMF with a proof-of-principle experiment which irradiated a naturally-abundant molybdenum foil at various proton beam energies. Delayed characteristic γ rays were measured with two Compton-shielded LaBr₃ scintillators. The technique was successfully demonstrated by relating the relative intensity of two γ-ray peaks to the energy of the beam at the Mo target, opening the door to future clinical applications where the range of the beam can be verified in real time.

GEANT4 simulation of a range verification method using delayed γ spectroscopy of a 92Mo marker

In this work, we propose a novel technique for in-vivo proton therapy range verification. This technique makes use of a molybdenum hadron tumour marker, implanted at a short distance from the clinical treatment volume. Signals emitted from the marker during treatment can provide a direct measurement of the proton beam energy at the marker's position. Fusion-evaporation reactions between the proton beam and marker nucleus result in the emission of delayed characteristic γ rays, which are detected off-beam for an improved signal-to-noise ratio. In order to determine the viability of this technique and to establish an experimental setup for future work, the Monte Carlo package GEANT4 was used in combination with ROOT to simulate a treatment scenario with the new method outlined in this work. These simulations show that the intensity of delayed γ rays produced from competing reactions yields a precise measurement of the range of the proton beam relative to the marker, with sub-millimetre uncertainty.

3D Compton image reconstruction method for whole gamma imaging

Compton imaging or Compton camera imaging has been studied well, but its advantages in nuclear medicine and molecular imaging have not been demonstrated yet. Therefore, the aim of this work was to compare Compton imaging with positron emission tomography (PET) by using the same imaging platform of whole gamma imaging (WGI). WGI is a concept that combines PET with Compton imaging by inserting a scatterer ring into a PET ring. This concept utilizes diverse types of gamma rays for 3D tomographic imaging. In this paper, we remodeled our previous WGI prototype for small animal imaging, and we developed an image reconstruction method based on a list-mode ordered subset expectation maximization algorithm incorporating detector response function modeling, random correction and normalization (sensitivity correction) for either PET and Compton imaging. To the best of our knowledge, this is the world's first realization of a full-ring Compton imaging system. We selected ⁸⁹Zr as an imaging target because a ⁸⁹Zr nuclide emits a 909 keV single-gamma ray as well as a positron, and we can directly compare Compton imaging of 909 keV photons with PET, a well-established modality. We measured a cylindrical phantom and a small rod phantom filled with ⁸⁹Zr solutions of 10.3 MBq and 10.2 MBq activity, respectively, for 1 h each. The uniform radioactivity distribution of the cylindrical phantom was reconstructed with normalization in both PET and Compton imaging. Coefficients of variation for region-of-interest values were 4.2% for Compton imaging and 3.3% for PET; the difference might be explained by the difference in the detected count number. The small rod phantom experiment showed that the WGI Compton imaging had spatial resolution better than 3.0 mm at the peripheral region although the center region had lower resolution. PET resolved 2.2 mm rods clearly at any location. We measured a mouse for 1 h, 1 d after injection of 9.8 MBq ⁸⁹Zr oxalate. The ⁸⁹Zr assimilated in the mouse bony structures was clearly depicted, and Compton imaging results agreed well with PET images, especially for the region inside the scatterer ring. In conclusion, we demonstrated the performance of WGI using the developed Compton image reconstruction method. We realized Compton imaging with a quality approaching that of PET, which is supporting a future expectation that Compton imaging outperforms PET.

Prompt gamma imaging for the identification of regional proton range deviations due to anatomic change in a heterogeneous region

OBJECTIVES

Prompt gamma (PG) imaging has previously been demonstrated for use in proton range verification of a brain treatment with a homogeneous target region. In this study, the feasibility of PG imaging to detect anatomic change within a heterogeneous region is presented.

METHODS

A prompt gamma camera recorded several fractions of a patient treatment to the base of skull. An evaluation CT revealed a decrease in sinus cavity filling during the treatment course. Comparison of PG profiles between measurement and simulation was performed to investigate range variations between planned and measured pencil beam spot positions.

RESULTS

For one field, an average over range of 3 mm due to the anatomic change could be detected for a subset of spots traversing the sinus cavity region. The two other fields appeared less impacted by the change but predicted range variations could not be detected. These results were partially consistent with the simulations of the evaluation CT.

CONCLUSION

We report the first clinical application of PG imaging that detected some of the expected small regional proton range deviations due to anatomic change in a heterogeneous region. However, several limitations exist with the technology that may limit its sensitivity to detect range deviations in heterogeneous regions.

ADVANCES IN KNOWLEDGE

We report on the first detection of range variations due to anatomic change in a heterogeneous region using PGI. The results confirm the feasibility of using PG-based range verification in highly heterogeneous target regions to identify deviations from the treatment plan.

Classification of the source of treatment deviation in proton therapy using prompt-gamma imaging information

PURPOSE

Prompt-gamma imaging (PGI)-based range verification has been successfully implemented in clinical proton therapy recently and its sensitivity to detect treatment deviations is currently investigated. The cause of treatment deviations can be multiple - for example, computed tomography (CT)-based range prediction, patient setup, and anatomical changes. Hence, it would be beneficial, if PGI-based verification would not only detect a treatment deviation but would also be able to directly identify its most probable source. Here, we propose a heuristically derived decision tree approach to automatically classify the sources of range deviation in proton pencil-beam scanning (PBS) treatments of head and neck tumors based on range information obtained with a PGI slit camera.

MATERIALS AND METHODS

The decision tree model was iteratively generated on a training dataset of different anatomical complexities, for an anthropomorphic head phantom and patient CT data (planning and control CTs) from five patients. Different range prediction errors, setup changes and relevant and nonrelevant anatomical changes were introduced or derived from control CTs, summing up to a total of 98 training scenarios. Independent validation was performed for another 98 scenarios, derived solely from patient CT data of another seven patients. PBS head and neck treatment plans were generated for the nominal scenario. For all PBS spots in the investigated field, PGI profiles were simulated using a dedicated analytical model of the slit camera for the nominal as well as the different error scenarios. From comparison of PGI profiles for nominal and error scenarios, a spot-wise range shift after spot aggregation with a kernel of 7 mm sigma was determined for each error scenario. The heuristic approach includes a prefiltering of the most suitable PBS spots for PGI treatment verification. From the validation, the accuracy, sensitivity, and specificity of the model were determined.

RESULTS

A five-step consecutive filter was developed to preselect PBS spots. On average, 25% of spots (1044 spots) remained as input for the classification model. The derived heuristic decision tree model is based on five parameters: The coefficient of determination (R² ), the slope and intercept of the linear regression between PGI-derived range shifts and the respectively predicted proton ranges for the investigated PBS spots, as well as the average and standard deviation of the PGI-derived shifts. With this approach, 94 of 98 error scenarios could be classified correctly in validation (accuracy of 96%). A sensitivity and specificity of 100% and 86% were reached.

CONCLUSIONS

In this simulation study it was demonstrated that the source of a treatment deviation can be identified from simulated noiseless PGI information in head and neck tumor treatments with high sensitivity and specificity. The application, refinement, and evaluation of the approach on measured PGI data will be the next step to show the clinical feasibility of PGI-based error source classification.

Computational model for detector timing effects in Compton-camera based prompt-gamma imaging for proton radiotherapy

This paper describes a realistic simulation of a Compton-camera (CC) based prompt-gamma (PG) imaging system for proton range verification for a range of clinical dose rates, and its comparison to PG measured data with a pre-clinical CC. We used a Monte Carlo plus Detector Effects (MCDE) model to simulate the production of prompt gamma-rays (PG) and their energy depositions in the CC. With Monte Carlo, we simulated PG emission resulting from irradiation of a high density polyethylene phantom with a 150 MeV proton pencil beam at dose rates of 5.0 × 10⁸, 2.6 × 10⁹, and 4.6 × 10⁹ p⁺ s^-1. Realistic detector timing effects (e.g. delayed triggering time, event-coincidence, dead time, etc,) were added in post-processing to allow for flexible count rate variations. We acquired PG emission measurements with our pre-clinical CC during irradiation with a clinical 150 MeV proton pencil beam at the same dose rates. For simulations and measurements, three primary changes could be seen in the PG emission data as the dose rate increased: (1) reduction in the total number of detected events due to increased dead-time percentage; (2) increase in false-coincidence events (i.e. multiple PGs interacting, rather than a single PG scatter); and (3) loss of distinct PG emission peaks in the energy spectrum. We used the MCDE model to estimate the quality of our measured PG data, primarily with regards to true and false double-scatters and triple-scatters recorded by the CC. The simulation results showed that of the recorded double-scatter PG interactions 22%, 57%, and 70% were false double-scatters and for triple-scatter interactions 3%, 21%, and 35% were false events at 5.0 × 10⁸, 2.6 × 10⁹, and 4.6 × 10⁹ p⁺ s^-1, respectively. These false scatter events represent noise in the data, and the high percentage of these events in the data represents a major limitation in our ability to produce usable PG images with our prototype CC.

Influence of momentum acceptance on range monitoring of 11C and 15O ion beams using in-beam PET

In heavy-ion therapy, the stopping position of primary ions in tumours needs to be monitored for effective treatment and to prevent overdose exposure to normal tissues. Positron-emitting ion beams, such as ¹¹C and ¹⁵O, have been suggested for range verification in heavy-ion therapy using in-beam positron emission tomography (PET) imaging, which offers the capability of visualizing the ion stopping position with a high signal-to-noise ratio. We have previously demonstrated the feasibility of in-beam PET imaging for the range verification of ¹¹C and ¹⁵O ion beams and observed a slight shift between the beam stopping position and the dose peak position in simulations, depending on the initial beam energy spread. In this study, we focused on the experimental confirmation of the shift between the Bragg peak position and the position of the maximum detected positron-emitting fragments via a PET system for positron-emitting ion beams of ¹¹C (210 MeV u^-1) and ¹⁵O (312 MeV u^-1) with momentum acceptances of 5% and 0.5%. For this purpose, we measured the depth doses and performed in-beam PET imaging using a polymethyl methacrylate (PMMA) phantom for both beams with different momentum acceptances. The shifts between the Bragg peak position and the PET peak position in an irradiated PMMA phantom for the ¹⁵O ion beams were 1.8 mm and 0.3 mm for momentum acceptances of 5% and 0.5%, respectively. The shifts between the positions of two peaks for the ¹¹C ion beam were 2.1 mm and 0.1 mm for momentum acceptances of 5% and 0.5%, respectively. We observed larger shifts between the Bragg peak and the PET peak positions for a momentum acceptance of 5% for both beams, which is consistent with the simulation results reported in our previous study. The biological doses were also estimated from the calculated relative biological effectiveness (RBE) values using a modified microdosimetric kinetic model (mMKM) and Monte Carlo simulation. Beams with a momentum acceptance of 5% should be used with caution for therapeutic applications to avoid extra dose to normal tissues beyond the tumour when the dose distal fall-off is located beyond the treatment volume.

Sensitivity improvement of YAP(Ce) cameras for imaging of secondary electron bremsstrahlung x-rays emitted during carbon-ion irradiation: problem and solution

Low-energy x-ray imaging of secondary electron bremsstrahlung x-rays emitted during carbon-ion irradiation is a promising method for range estimation and could be used for imaging with almost clinical dose levels of carbon ion. However, the number of counts in images with clinical dose levels is relatively small, making it difficult to obtain precise range estimations. Since improving the sensitivity of the x-ray camera may solve this issue, we developed two new types of x-ray cameras. One uses a 1 mm thick, 40 mm × 40 mm cerium-doped yttrium aluminum perovskite (YA1O₃: YAP(Ce)) scintillator plate combined with a 2 inch square flat panel photomultiplier tube (FP-PMT) contained in a 2 cm thick tungsten shield with a pinhole collimator positioned 50 mm from the scintillator; the other uses a 0.5 mm thick, 20 mm × 20 mm YAP(Ce) scintillator plate combined with a 1 inch square position sensitive photomultiplier tube (PSPMT) contained in the same tungsten shield with a pinhole collimator, but with the scintillator positioned closer (30 mm) to the pinhole collimator to obtain a similar field of view. For both cameras, we used a wider angle (∼55°) pinhole collimator to measure the phantom closer to improve sensitivity. Although the 40 mm × 40 mm YAP(Ce) camera had high system spatial resolution, the background count fractions were high and produced a high count area at the center of the images due to the pulse pileup of the signals. With the 20 mm × 20 mm YAP(Ce) camera, we obtained x-ray images with low background counts without a high count area at the image center. By smoothing the measured images, we were able to estimate the ranges even for clinical dose levels. We therefore confirmed that one of our newly developed YAP(Ce) cameras had high sensitivity and is promising for the imaging of secondary electron bremsstrahlung x-rays during irradiation of carbon ions in clinical conditions.

A filtering approach for PET and PG predictions in a proton treatment planning system

Positron emission tomography (PET) and prompt gamma (PG) detection are promising proton therapy monitoring modalities. Fast calculation of the expected distributions is desirable for comparison to measurements and to develop/train algorithms for automatic treatment error detection. A filtering formalism was used for positron-emitter predictions and adapted to allow for its use for the beamline of any proton therapy centre. A novel approach based on a filtering formalism was developed for the prediction of energy-resolved PG distributions for arbitrary tissues. The method estimates PG yields and their energy spectra in the entire treatment field. Both approaches were implemented in a research version of the RayStation treatment planning system. The method was validated against PET monitoring data and Monte Carlo simulations for four patients treated with scanned proton beams. Longitudinal shifts between profiles from analytical and Monte Carlo calculations were within -1.7 and 0.9 mm, with maximum standard deviation of 0.9 mm and 1.1 mm, for positron-emitters and PG shifts, respectively. Normalized mean absolute errors were within 1.2 and 5.3%. When comparing measured and predicted PET data, the same more complex case yielded an average shift of 3 mm, while all other cases were below absolute average shifts of 1.1 mm. Normalized mean absolute errors were below 7.2% for all cases. A novel solution to predict positron-emitter and PG distributions in a treatment planning system is proposed, enabling calculation times of only a few seconds to minutes for entire patient cases, which is suitable for integration in daily clinical routine.

PIBS: Proton and ion beam spectroscopy for in vivo measurements of oxygen, carbon, and calcium concentrations in the human body

Proton and ion beam therapy has proven to benefit tumour control with lower side-effects, mostly in paediatrics. Here we demonstrate a feasible technique for proton and ion beam spectroscopy (PIBS) capable of determining the elemental compositions of the irradiated tissues during particle therapy. This follows the developments in prompt gamma imaging for online range verification and the inheritance from prompt gamma neutron activation analysis. Samples of water solutions were prepared to emulate varying oxygen and carbon concentrations. The irradiation of those samples and other tissue surrogate inserts by protons and ion beams under clinical conditions clearly showed a logarithmic relationship between the target elemental composition and the prompt gamma production. This finding is in line with the known logarithmic dependence of the pH with the proton molar concentration. Elemental concentration changes of 1% for calcium and 2% for oxygen in adipose, brain, breast, liver, muscle and bone-related tissue surrogates were clearly identified. Real-time in vivo measurements of oxygen, carbon and calcium concentrations will be evaluated in a pre-clinical and clinical environment. This technique should have an important impact in the assessment of tumour hypoxia over the course of several treatment fractions and the tracking of calcifications in brain metastases.

Latest developments in in-vivo imaging for proton therapy

Owing to the favorable physical and biological properties of swift ions in matter, their application to radiation therapy for highly selective cancer treatment is rapidly spreading worldwide. To date, over 90 ion therapy facilities are operational, predominantly with proton beams, and about the same amount is under construction or planning.Over the last decades, considerable developments have been achieved in accelerator technology, beam delivery and medical physics to enhance conformation of the dose delivery to complex shaped tumor volumes, with excellent sparing of surrounding normal tissue and critical organs. Nevertheless, full clinical exploitation of the ion beam advantages is still challenged, especially by uncertainties in the knowledge of the beam range in the actual patient anatomy during the fractionated course of treatment, thus calling for continued multidisciplinary research in this rapidly emerging field.This contribution will review latest developments aiming to image the patient with the same beam quality as for therapy prior to treatment, and to visualize in-vivo the treatment delivery by exploiting irradiation-induced physical emissions, with different level of maturity from proof-of-concept studies in phantoms and first in-silico studies up to clinical testing and initial clinical evaluation.