On the use of machine learning methods for mPSD calibration in HDR brachytherapy.

We sought to evaluate the feasibility of using machine learning (ML) algorithms for multipoint plastic scintillator detector (mPSD) calibration in high-dose-rate (HDR) brachytherapy. Dose measurements were conducted under HDR brachytherapy conditions. The dosimetry system consisted of an optimized 1-mm-core mPSD and a compact assembly of photomultiplier tubes coupled with dichroic mirrors and filters. An 192Ir source was remotely controlled and sent to various positions in a homemade PMMA holder, ensuring 0.1-mm positional accuracy. Dose measurements covering a range of 0.5 to 12 cm of source displacement were carried out according to TG-43 U1 recommendations. Individual scintillator doses were decoupled using a linear regression model, a random forest estimator, and artificial neural network algorithms. The dose predicted by the TG-43U1 formalism was used as the reference for system calibration and ML algorithm training. The performance of the different algorithms was evaluated using different sample sizes and distances to the source for the mPSD system calibration. We found that the calibration conditions influenced the accuracy in predicting the measured dose. The decoupling methods' deviations from the expected TG-43 U1 dose generally remained below 20%. However, the dose prediction with the three algorithms was accurate to within 7% relative to the dose predicted by the TG-43 U1 formalism when measurements were performed in the same range of distances used for calibration. In such cases, the predictions with random forest exhibited minimal deviations (<2%). However, the performance random forest was compromised when the predictions were done beyond the range of distances used for calibration. Because the linear regression algorithm can extrapolate the data, the dose prediction by the linear regression was less influenced by the calibration conditions than random forest. The linear regression algorithm's behavior along the distances to the source was smoother than those for the random forest and neural network algorithms, but the observed deviations were more significant than those for the neural network and random forest algorithms. The number of available measurements for training purposes influenced the random forest and neural network models the most. Their accuracy tended to converge toward deviation values close to 1% from a number of dwell positions greater than 100. In performing HDR brachytherapy dose measurements with an optimized mPSD system, ML algorithms are good alternatives for precise dose reporting and treatment assessment during this kind of cancer treatment.

3D source tracking and error detection in HDR using two independent scintillator dosimetry systems.

PURPOSE: The aim of this study is to perform three-dimensional (3D) source position reconstruction by combining in vivo dosimetry measurements from two independent detector systems. METHODS: Time-resolved dosimetry was performed in a water phantom during HDR brachytherapy irradiation with 192 Ir source using two detector systems. The first was based on three plastic scintillator detectors and the second on a single inorganic crystal (CsI:Tl). Brachytherapy treatments were simulated in water under TG-43U1 conditions, including a HDR prostate plan. Treatment needles were placed in distances covering a range of source movement of 120 mm around the detectors. The distance from each dwell position to each scintillator was determined based on the measured dose rates. The three distances given by the mPSD were recalculated to a position along the catheter (z) and a distance radially away from the mPSD (xy) for each dwell position (a circumference around the mPSD). The source x, y, and z coordinates were derived from the intersection of the mPSD's circumference with the sphere around the ISD based on the distance to this detector. We evaluated the accuracy of the source position reconstruction as a function of the distance to the source, the most likely location for detector positioning within a prostate volume, as well as the capacity to detect positioning errors. RESULTS: Approximately 4000 source dwell positions were tracked for eight different HDR plans. An intersection of the mPSD torus and the ISD sphere was observed in 77.2% of the dwell positions, assuming no uncertainty in the dose rate determined distance. This increased to 100% if 1σ search regions were added. However, only 73(96)% of the expected dwell positions were found within the intersection band for 1(2) σ uncertainties. The agreement between the source's reconstructed and expected positions was within 3 mm for a range of distances to the source up to 50 mm. The experiments on a HDR prostate plan, showed that by having at least one of the detectors located in the middle of the prostate volume, reduces the measurement deviations considerably compared to scenarios where the detectors were located outside of the prostate volume. The analysis showed a detection probability that, in most cases, is far from the random detection threshold. Errors of 1(2) mm can be detected in ranges of 5-25 (25-50) mm from the source, with a true detection probability rate higher than 80%, while the false probability rate is kept below 20%. CONCLUSIONS: By combining two detector responses, we enabled the determination of the absolute source coordinates. The combination of the mPSD and the ISD in vivo dosimetry constitutes a promising alternative for real-time 3D source tracking in HDR brachytherapy.

Monte Carlo dosimetric characterization of a new high dose rate 169 Yb brachytherapy source and independent verification using a multipoint plastic scintillator detector.

PURPOSE: A prototype 169 Yb source was developed in combination with a dynamic rotating platinum shield system (AIM-Brachy) to deliver intensity modulated brachytherapy (IMBT). The purpose of this study was to evaluate the dosimetric characteristics of the bare/shielded 169 Yb source using Monte Carlo (MC) simulations and perform an independent dose verification using a dosimetry platform based on a multipoint plastic scintillator detector (mPSD). METHODS: The TG-43U1 dosimetric parameters were calculated for the source model using RapidBrachyMCTPS. Real-time dose rate measurements were performed in a water tank for both the bare/shielded source using a custom remote afterloader. For each dwell position, the dose rate was independently measured by the three scintillators (BCF-10, BCF-12, and BCF-60). For the bare source, dose rate was measured at distances up to 3 cm away from the source over a range of 7 cm along the catheter. For the shielded source, measurements were performed with the mPSD placed at 1 cm from the source at four different azimuthal angles ( 0 ∘ , 9 0 ∘ , 18 0 ∘ , and 27 0 ∘ ). RESULTS: The dosimetric parameters were tabulated for the source model. For the bare source, differences between measured and calculated along-away dose rates were generally below 5-10%. Along the transverse axis, deviations were, on average (range), 3.3% (0.6-6.2%) for BCF-10, 1.7% (0.9-2.9%) for BCF-12, and 2.2% (0.3-4.4%) for BCF-60. The maximum dose rate reduction due to shielding at a radial distance of 1 cm was 88.8 ± 1.2%, compared to 83.5 ± 0.5% as calculated by MC. CONCLUSIONS: The dose distribution for the bare/shielded 169 Yb source was independently verified using mPSD with good agreement in regions close to the source. The 169 Yb source coupled with the partial-shielding system is an effective technique to deliver IMBT.

Dosimetric performance of a multipoint plastic scintillator dosimeter as a tool for real-time source tracking in high dose rate 192 Ir brachytherapy.

PURPOSE: This study aims to present the performance of a multipoint plastic scintillation detector (mPSD) as a tool for real-time dose measurements (covering three orders of magnitude in dose rate), source-position triangulation, and dwell time assessment in high dose rate (HDR) brachytherapy. METHODS: A previously characterized and optimized three-point sensor system was used for HDR brachytherapy measurements. The detector was composed of three scintillators: BCF60, BCF12, and BCF10. Scintillation light was transmitted through a single 1-mm-diameter clear optical fiber and read by a compact assembly of photomultiplier tubes (PMTs). Each component was numerically optimized to allow for signal deconvolution using a multispectral approach, taking care of the Cerenkov stem effect as well as extracting the dose from each scintillator. The PMTs were read simultaneously using a data acquisition board at a rate of 100 KHz and controlled with in-house software based on Python. An 192 I r source (Flexitron, Elekta-Brachy) was remotely controlled and sent to various positions in a in-house PMMA holder, ensuring 0.1 mm positional accuracy. Dose measurements covering a range of 10 cm of source movement were carried out according to TG-43 U1 recommendations. Water measurements were performed in order to: (a) characterize the system's response in terms of angular dependence; (b) obtain the relative contribution of positioning and measurement uncertainties to the total system uncertainty; (c) assess the system's temporal resolution; and (d) track the source position in real time. The triangulation principle was applied to report the source position in three-dimensional space. RESULTS: As expected, the positioning uncertainty dominated close to the source, whereas the measurement uncertainty dominated at larger distances. A maximum measurement uncertainty of 17 % was observed for the BCF60 scintillator at 10 cm from the source. Based on the uncertainty chain, the best compromises between positioning and measurement uncertainties were reached at 17.2, 17.4, and 17.5 mm for the BCF10, BCF12, and BCF60 scintillators, respectively, which also corresponded to the recommended optimal distances to the source for calibration purposes. The detector further exhibited no angular dependence. All dose values were found to be within 2% of the dose value at 90 ∘ . In the experiments performed for source-position determination, the system provided an average location with a standard deviation under 1.7 mm. The maximum observed differences between measured and expected values were 1.82 and 1.8 mm in the x- and z-directions, respectively. Deviations between the mPSD measurements and expected TG-43 values were below 5% in all the explored measurement conditions. With regard to dwell time measurement accuracy, the maximum deviation observed at all distances was 0.56 ± 0.25 s, with a weighted average of the three scintillators below 0.33 ± 0.37 s at all distances covered in this study. CONCLUSIONS: Real-time HDR brachytherapy measurements were performed with an optimized mPSD system. The performance of the system demonstrated that it could be used for simultaneous, in vivo, real-time reporting of dose, dwell time, and source position during HDR brachytherapy.

Optimization of a multipoint plastic scintillator dosimeter for high dose rate brachytherapy.

PURPOSE: This study is devoted to optimizing and characterizing the response of a multipoint plastic scintillator detector (mPSD) for application to in vivo dosimetry in high dose rate (HDR) brachytherapy. METHODS: An exhaustive analysis was carried out in order to obtain an optimized mPSD design that maximizes the scintillation light collection produced by the interaction of ionizing photons. More than 20 prototypes of mPSD were built and tested in order to determine the appropriate order of scintillators relative to the photodetector (distal, center, or proximal) as well as their length as a function of the scintillation light emitted. The available detecting elements are the BCF-60, BCF-12, and BCF-10 scintillators (Saint Gobain Crystals, Hiram, OH, USA), separated from each other by segments of Eska GH-4001 clear optical fibers (Mitsubishi Rayon Co., Ltd., Tokyo, Japan). The contribution of each scintillator to the total spectrum was determined by irradiations in the low energy range (<120 keV). For the best mPSD design, a numerical optimization was done in order to select the optical components [dichroic mirrors, filters, and photomultipliers tubes (PMTs)] that best match the light emission profile. Calculations were performed taking into account the measured scintillation spectrum and light yield, the manufacturer-reported transmission and attenuation of the optical components, and the experimentally characterized PMT noise. The optimized dosimetric system was used for HDR brachytherapy measurements. The system was independently controlled from the 192 Ir source via LabVIEW and read simultaneously using an NI-DAQ board. Dose measurements as a function of distance from the source were carried out according to TG-43U1 recommendations. The system performance was quantified in terms of signal to noise ratio (SNR) and signal to background ratio (SBR). RESULTS: For best overall light-yield emission, it was determined that BCF-60 should be placed at the distal position, BCF-12 in the center, and BCF-10 at the proximal position with respect to the photodetector. This configuration allowed for optimized light transmission through the collecting fiber and avoided inter-scintillator excitation and self-absorption effects. The optimal scintillator length found was of 3, 6, and 7 mm for BCF-10, BCF- 12, and BCF-60, respectively. The optimized luminescence system allowed for signal deconvolution using a multispectral approach, extracting the dose to each element while taking into account the Cerenkov stem effect. Differences between the mPSD measurements and TG-43U1 remain below 5% in the range of 0.5 to 6.5 cm from the source. The dosimetric system can properly differentiate the scintillation signal from the background for a wide range of dose rate conditions; the SNR was found to be above 5 for dose rates above 22 mGy/s while the minimum SBR measured was 1.8 at 6 mGy/s. CONCLUSION: Based on the spectral response at different conditions, an mPSD was constructed and optimized for HDR brachytherapy dosimetry. It is sensitive enough to allow multiple simultaneous measurements over a clinically useful distance range, up to 6.5 cm from the source. This study constitutes a baseline for future applications enabling real-time dose measurements and source position reporting over a wide range of dose rate conditions.

Aplicación de un modelo de un acelerador Elekta Precise basado en Monte Carlo para la evaluación de la respuesta de detectores de estado sólido / Application of an Elekta Precise lineal accelerator model based on Monte Carlo to evaluate the solid state detectors response

En el trabajo se realizó el análisis de la respuesta de diferentes detectores de estado sólido a partir de un modelo de Monte Carlo de un acelerador lineal Elekta Precise, para los haces de energías de 6 MV y 15 MV. Para ello se realizaron simulaciones con el código EGSnrc. Se calculó la dosis depositada en un maniquí de agua voxelizado con su superficie a 100 cm de la fuente, empleando los valoresóptimos de energía media y FWHM del haz primario de electrones para este modelo. A partir de la dosis depositada en el maniquí se construyeron las curvas de dosis en profundidad y perfiles de dosis a diferentes profundidades. Las curvas se compararon con valores medidos para cada detector empleado en un arreglo experimental similar a la simulación realizada, aplicando criterios de aceptabilidad basados en intervalos de confianza. De forma adicional se analizó para cada caso la dosis en función del tamaño de campo. Se obtuvo una buena correspondencia entre las simulaciones y las mediciones, encontrándose todos los resultados dentro de los márgenes de tolerancia.

The evaluation of the solid state detectors response based on a Monte Carlo model of an Elekta Precise lineal accelerator, was done in this work for the beam energies of 6 y 15 MV. Simulations were performed using the EGSnrc code. Employing the optimal values of mean energy and FWHM from the primary electron beam, deposited dose in a voxelized water phantom at 100 cm of source to surface distance was calculated. Depth dose curves and lateral dose profi les were obtained. Comparison between simulations and the experimental values obtained for each detector, were done using acceptability criteria based on confi dence limits. Additionally outputs factors were analyzed in each one of the study cases. Good agreement between simulations and measurements were reached.

Accurate model of photon beams as a tool for commissioning and quality assurance of treatment planning calculations / Modelo de haces de fotones como una herramienta para el control de calidad de los cálculos en la planificación de tratamientos

Simulation of a linear accelerator (linac) head requires determining the parameters that characterize the primary electron beam striking on the target which is a step that plays a vital role in the accuracy of Monte Carlo calculations. In this work, the commissioning of photon beams (6 MV and 15 MV) of an Elekta Precise accelerator, using the Monte Carlo code EGSnrc, was performed. The influence of the primary electron beam characteristics on the absorbed dose distribution for two photon qualities was studied. Using different combinations of mean energy and radial FWHM of the primary electron beam, deposited doses were calculated in a water phantom, for different field sizes. Based on the deposited dose in the phantom, depth dose curves and lateral dose profiles were constructed and compared with experimental values measured in an arrangement similar to the simulation. Taking into account the main differences between calculations and measurements, an acceptability criteria based on confidence limits was implemented. As expected, the lateral dose profiles for small field sizes were strongly infl uenced by the radial distribution (FWHM). The combinations of energy/FWHM that best reproduced the experimental results were used to generate the phase spaces, in order to obtain a model with the motorized wedge included and to calculate output factors. A good agreement was obtained between simulations and measurements for a wide range of fi eld sizes , being all the results found within the range of tolerance.

La simulación del cabezal de un acelerador lineal requiere de la determinación de los parámetros que caracterizan el haz primario de electrones que incide en el blanco de radiación, los cuales juegan un papel importante en la exactitud de los cálculos con Monte Carlo. En este trabajo se realizó la habilitación de los haces de fotones (6 MV y 15 MV) de un acelerador Elekta Precise, empleando el código de Monte Carlo EGSnrc. De forma adicional se estudió la influencia que ejerce cambios en las características del haz primario de electrones sobre la distribución de dosis absorbida en diferentes campos de radiación. Basado en la dosis absorbida, curvas de dosis en profundidad y perfiles de dosis se calcularon y compararon con valores experimentales medidos en un arreglo similar a las simulaciones empleando criterios de aceptabilidad. Los perfiles de dosis para campos pequeños resultaron ser fuertemente dependientes de la distribución radial (FWHM). Las combinaciones de energías/FWHM que mejor se ajustaron a las mediciones se emplearon en la generación de espacios de fases, para obtener un modelo con la cuña motorizada y para el cálculo de los factores de campo. Se obtuvo muy buena correspondencia entre las mediciones y las simulaciones realizadas, encontrándose todos los resultados dentro de los márgenes de tolerancias.