High-resolution entry and exit surface dosimetry in a 1.5 T MR-linac

The magnetic field of a transverse MR-linac alters electron trajectories as the photon beam transits through materials, causing lower doses at flat entry surfaces and increased doses at flat beam-exiting surfaces. This study investigated the response of a MOSFET detector, known as the MOSkin™, for high-resolution surface and near-surface percentage depth dose measurements on an Elekta Unity. Simulations with Geant4 and the Monaco treatment planning system (TPS), and EBT-3 film measurements, were also performed for comparison. Measured MOSkin™ entry surface doses, relative to D_max, were (9.9 ± 0.2)%, (10.1 ± 0.3)%, (11.3 ± 0.6)%, (12.9 ± 1.0)%, and (13.4 ± 1.0)% for 1 × 1 cm², 3 × 3 cm², 5 × 5 cm², 10 × 10 cm², and 22 × 22 cm² fields, respectively. For the investigated fields, the maximum percent differences of Geant4, TPS, and film doses extrapolated and interpolated to a depth suitable for skin dose assessment at the beam entry, relative to MOSkin™ measurements at an equivalent depth were 1.0%, 2.8%, and 14.3%, respectively, and at a WED of 199.67 mm at the beam exit, 3.2%, 3.7% and 5.7%, respectively. The largest measured increase in exit dose, due to the electron return effect, was 15.4% for the 10 × 10 cm² field size using the MOSkin™ and 17.9% for the 22 × 22 cm² field size, using Geant4 calculations. The results presented in the study validate the suitability of the MOSkin™ detector for transverse MR-linac surface dosimetry.

Small beams, fast predictions: a comparison of machine learning dose prediction models for proton minibeam therapy

BACKGROUND

Dose calculations for novel radiotherapy cancer treatments such as proton minibeam radiation therapy is often done using full Monte Carlo (MC) simulations. As MC simulations can be very time consuming for this kind of application, deep learning models have been considered to accelerate dose estimation in cancer patients.

PURPOSE

This work systematically evaluates the dose prediction accuracy, speed and generalization performance of three selected state-of-the-art deep learning models for dose prediction applied to the proton minibeam therapy. The strengths and weaknesses of those models are thoroughly investigated, helping other researchers to decide on a viable algorithm for their own application.

METHODS

The following recently published models are compared: first, a 3D U-Net model trained as a regression network, second, a 3D U-Net trained as a generator of a generative adversarial network (GAN) and third, a dose transformer model which interprets the dose prediction as a sequence translation task. These models are trained to emulate the result of MC simulations. The dose depositions of a proton minibeam with a diameter of 800μm and an energy of 20-100 MeV inside a simple head phantom calculated by full Geant4 MC simulations are used as a case study for this comparison. The spatial resolution is 0.5 mm. Special attention is put on the evaluation of the generalization performance of the investigated models.

RESULTS

Dose predictions with all models are produced in the order of a second on a GPU, the 3D U-Net models being fastest with an average of 130 ms. An investigated 3D U-Net regression model is found to show the strongest performance with overall 61.0 % ± $\%\pm$ 0.5% of all voxels exhibiting a deviation in energy deposition prediction of less than 3% compared to full MC simulations with no spatial deviation allowed. The 3D U-Net models are observed to show better generalization performance for target geometry variations, while the transformer-based model shows better generalization with regard to the proton energy.

CONCLUSIONS

This paper reveals that (1) all studied deep learning models are significantly faster than non-machine learning approaches predicting the dose in the order of seconds compared to hours for MC, (2) all models provide reasonable accuracy, and (3) the regression-trained 3D U-Net provides the most accurate predictions.

4He dose- and track-averaged linear energy transfer: Monte Carlo algorithms and experimental verification

Objective. In the present hadrontherapy scenario, there is a growing interest in exploring the capabilities of different ion species other than protons and carbons. The possibility of using different ions paves the way for new radiotherapy approaches, such as the multi-ions treatment, where radiation could vary according to target volume, shape, depth and histologic characteristics of the tumor. For these reasons, in this paper, the study and understanding of biological-relevant quantities was extended for the case of 4He ion. Approach. Geant4 Monte Carlo based algorithms for dose- and track-averaged LET (Linear Energy Transfer) calculations, were validated for 4He ions and for the case of a mixed field characterised by the presence of secondary ions from both target and projectile fragmentation. The simulated dose and track averaged LETs were compared with the corresponding dose and frequency mean values of the lineal energy, y D ¯ and y ¯ F , derived from experimental microdosimetric spectra. Two microdosimetric experimental campaigns were carried out at the Italian eye proton therapy facility of the Laboratori Nazionali del Sud of Istituto Nazionale di Fisica Nucleare (INFN-LNS, Catania, I) using two different microdosimeters: the MicroPlus probe and the nano-TEPC (Tissue Equivalent Proportional Counter). Main results. A good agreement of L ¯ d Total and L ¯ t Total with y ¯ D and y ¯ T experimentally measured with both microdosimetric detectors MicroPlus and nano-TEPC in two configurations: full energy and modulated 4He ion beam, was found. Significance. The results of this study certify the use of a very effective tool for the precise calculation of LET, given by a Monte Carlo approach which has the advantage of allowing detailed simulation and tracking of nuclear interactions, even in complex clinical scenarios.

Characterisation of a well-type NaI(Tl) detector by means of a Monte Carlo simulation for radionuclide metrology application

A well-type NaI(Tl) detector was modelled and characterised by means of a Monte Carlo simulation, as part of a project to develop a 4πβ (Plastic Scintillator)-4πγ instrument to be used for the primary standardisation of radionuclides at ANSTO. The simulation based on GEANT4 was used to characterise the 4πγ detector in terms of potential dead layer/inactive materials, full energy peak efficiency, coincidence-summing correction, and energy resolution. An excellent agreement was obtained between the simulation results and the experimental measurements.

Solid-phase combinatorial synthesis using microarrays of microcompartments with light-induced on-chip cell screening

The process of drug discovery includes individual synthesis and characterization of drug candidates, followed by a biological screening, which is separated from synthesis in space and time. This approach suffers from low throughput and associated high costs, which in turn lead to inefficiency in the field of drug discovery. Here, we present a miniaturized platform combining combinatorial solid-phase synthesis with high-throughput cell screenings. The method is based on the formation of nanoporous poly(2-hydroxyethyl methacrylate-co-ethylene dimethacrylate) layers patterned with hydrophilic spots separated from each other by superhydrophobic liquid-impermeable barriers. The porous polymer inside the hydrophilic spots is used as a support to conduct solid-phase synthesis. The hydrophilic spots can be then filled with droplets containing either reagents for synthesis or live cells. Upon irradiation with UV light, products of solid-phase synthesis are released from the porous polymer because of the photo-cleavable linkers used and diffuse into separate droplets. The light-induced release of the products allows the control of the release spatially, temporally, and quantitatively. To demonstrate the versatility and usability of the platform for various cell lines, we have successfully implemented peptide synthesis to create an exemplary chemical library and demonstrated high cell viability after the UV-triggered small-molecule release.

High toxicity of Bi(OH)3 and α-Bi2O3 nanoparticles towards malignant 9L and MCF-7 cells

Here we report the extreme toxicity in vitro of Bi(OH)₃ and α-Bi₂O₃ nanoparticles (NPs), obtained through a facile synthesis with an average single particle size of 6-10 nm, tested on malignant gliosarcoma 9L and MCF-7 human breast cancer cells. For both nanomaterials, clonogenic assays reveal a mortality of over 90% in 9L and MCF-7 cells for a concentration of 50 μg/mL after incubation for 24 h. Moreover, the NPs show a significant mortality of up to 60% in the malignant cells at the very low concentration of 6.25 μg/mL. In contrast, at the same concentration, the nanomaterials exhibit no noticeable mortality towards normal Madin-Darby canine kidney cells. The internalisation of the NPs was demonstrated using flow cytometry and confocal microscopy was used to investigate when the loss of cell viability starts. The NPs show a faster cell death in 9L cells compared with MCF-7 cells, demonstrated via the identification of apoptosis through increased sub G₁ levels after 24 h of NP incubation. Cleavage is identified as the main apoptotic nuclear morphology in 9L, which suggests the presence of reactive oxygen species.

In vivo dosimetry: trends and prospects for brachytherapy

The error types during brachytherapy (BT) treatments and their occurrence rates are not well known. The limited knowledge is partly attributed to the lack of independent verification systems of the treatment progression in the clinical workflow routine. Within the field of in vivo dosimetry (IVD), it is established that real-time IVD can provide efficient error detection and treatment verification. However, it is also recognized that widespread implementations are hampered by the lack of available high-accuracy IVD systems that are straightforward for the clinical staff to use. This article highlights the capabilities of the state-of-the-art IVD technology in the context of error detection and quality assurance (QA) and discusses related prospects of the latest developments within the field. The article emphasizes the main challenges responsible for the limited practice of IVD and provides descriptions on how they can be overcome. Finally, the article suggests a framework for collaborations between BT clinics that implemented IVD on a routine basis and postulates that such collaborations could improve BT QA measures and the knowledge about BT error types and their occurrence rates.

Online in vivo dosimetry in high dose rate prostate brchytherapy with MOSkin detectors: in phantom feasibility study

MOSkin detectors were studied to perform real-time in vivo dose measurements in high dose rate prostate brachytherapy. Measurements were performed inside an urethral catheter in a gel phantom simulating a real prostate implant. Measured and expected doses were compared and the discrepancy was found to be within 8.9% and 3.8% for single MOSkin and dual-MOSkin configurations, respectively. Results show that dual-MOSkin detectors can be profitably adopted in prostate brachytherapy treatments to perform real-time in vivo dosimetry inside the urethra.

Alpha particle and proton relative thermoluminescence efficiencies in LiF:Mg,Cu,P:is track structure theory up to the task?

Low-energy alpha particle and proton heavy charged particle (HCP) relative thermoluminescence (TL) efficiencies are calculated for the major dosimetric glow peak in LiF:Mg,Cu,P (MCP-N) in the framework of track structure theory (TST). The calculations employ previously published TRIPOS-E Monte Carlo track segment values of the radial dose in condensed phase LiF calculated at the Instituto National de Investigaciones Nucleares (Mexico) and experimentally measured normalised (60)Co gamma-induced TL dose-response functions, f(D), carried out at the Institute of Nuclear Physics (Poland). The motivation for the calculations is to test the validity of TST in a TL system in which f(D) is not supralinear (f(D) >1) and is not significantly dependent on photon energy contrary to the behaviour of the dose-response of composite peak 5 in the glow curve of LiF:Mg,Ti (TLD-100). The calculated HCP relative efficiencies in LiF:MCP-N are 23-87% lower than the experimentally measured values, indicating a weakness in the major premise of TST which exclusively relates HCP effects to the radiation action of the secondary electrons liberated by the HCP slowing down. However, an analysis of the uncertainties involved in the TST calculations and experiments (i.e. experimental measurement of f(D) at high levels of dose, sample light self-absorption and accuracy in the estimation of D(r), especially towards the end of the HCP track) indicate that these may be too large to enable a definite conclusion. More accurate estimation of sample light self-absorption, improved measurements of f(D) and full-track Monte Carlo calculations of D(r) incorporating improvements of the low-energy electron transport are indicated in order to reduce uncertainties and enable a final conclusion.

Mysteries of LiF TLD response following high ionisation density irradiation: nanodosimetry and track structure theory, dose response and glow curve shapes

Three outstanding effects of ionisation density on the thermoluminescence (TL) mechanisms giving rise to the glow peaks of LiF:Mg,Ti (TLD-100) are currently under investigation: (1) the dependence of the heavy charged particle (HCP) relative efficiency with increasing ionisation density and the effectiveness of its modelling by track structure theory (TST), (2) the behaviour of the TL efficiency, f(D), as a function of photon energy and dose. These studies are intended to promote the development of a firm theoretical basis for the evaluation of relative TL efficiencies to assist in their application in mixed radiation fields. And (3) the shape of composite peak 5 in the glow curve for various HCP types and energies and following high-dose electron irradiation, i.e. the ratio of the intensity of peak 5a to peak 5. Peak 5a is a low-temperature satellite of peak 5 arising from electron-hole capture in a spatially correlated trapping centre/luminescent centre (TC/LC) complex that has been suggested to possess a potential as a solid-state nanodosemeter due to the preferential electron/hole population of the TC/LC at high ionisation density. It is concluded that (1) the predictions of TST are very strongly dependent on the choice of photon energy used in the determination of f(D); (2) modified TST employing calculated values of f(D) at 2 keV is in agreement with 5-MeV alpha particle experimental results for composite peak 5 but underestimates the 1.5-MeV proton relative efficiencies. Both the proton and alpha particle relative TL efficiencies of the high-temperature TL (HTTL) peaks 7 and 8 are underestimated by an order of magnitude suggesting that the HTTL efficiencies are affected by other factors in addition to radial electron dose; (3) the dose-response supralinearity of peaks 7 and 8 change rapidly with photon energy: this behaviour is explained in the framework of the unified interaction model as due to a very strong dependence on photon energy of the relative intensity of localised recombination and (4) the increased width and decrease in T(max) of composite peak 5 as a function of ionisation density is due to the greater relative intensity of peak 5a (a low-temperature component of peak 5 arising from two-energy transfer events, which leads to localised recombination).

Thermoluminescence solid-state nanodosimetry--the peak 5A/5 dosemeter

The shape of composite peak 5 in the glow curve of LiF:Mg,Ti (TLD-100) following (90)Sr/(90)Y beta irradiation, previously demonstrated to be dependent on the cooling rate used in the 400°C pre-irradiation anneal, is shown to be dependent on ionisation density in both naturally cooled and slow-cooled samples. Following heavy-charged particle high-ionisation density (HID) irradiation, the temperature of composite peak 5 decreases by ∼5°C and the peak becomes broader. This behaviour is attributed to an increase in the relative intensity of peak 5a (a low-temperature satellite of peak 5). The relative intensity of peak 5a is estimated using a computerised glow curve deconvolution code based on first-order kinetics. The analysis uses kinetic parameters for peaks 4 and 5 determined from ancillary measurements resulting in nearly 'single-glow peak' curves for both the peaks. In the slow-cooled samples, owing to the increased relative intensity of peak 5a compared with the naturally cooled samples, the precision of the measurement of the 5a/5 intensity ratio is found to be ∼15% (1 SD) compared with ∼25% for the naturally cooled samples. The ratio of peak 5a/5 in the slow-cooled samples is found to increase systematically and gradually through a variety of radiation fields from a minimum value of 0.13±0.02 for (90)Sr/(90)Y low-ionisation density irradiations to a maximum value of ∼0.8 for 20 MeV Cu and I ion HID irradiations. Irradiation by low-energy electrons of energy 0.1-1.5 keV results in values between 1.27 and 0.95, respectively. The increasing values of the ratio of peak 5a/5 with increasing ionisation density demonstrate the viability of the concept of the peak 5a/5 nanodosemeter and its potential in the measurement of average ionisation density in a 'nanoscopic' mass containing the trapping centre/luminescent centre spatially correlated molecule giving rise to composite peak 5.

Single-pulse ultrafast laser imprinting of axial dot arrays in bulk glasses

Ultrafast laser processing of bulk transparent materials can significantly gain flexibility when the number of machining spots is increased. We present a photoinscription regime in which an array of regular dots is generated before the region of main laser focus under single-pulse exposure in fused silica and borosilicate crown glass without any external spatial phase modulation. The specific position of the dots does not rely on nonlinear propagation effects but is mainly determined by beam truncation and is explained by a Fresnel propagation formalism taking into account beam apodization and linear wavefront distortions at the air/glass interface. The photoinscription regime is employed to generate a two-dimensional array of dots in fused silica. We show that an additional phase modulation renders flexible the pattern geometry.

Nanosize structural modifications with polarization functions in ultrafast laser irradiated bulk fused silica

Laser-induced self-organization of regular nanoscale layered patterns in fused silica is investigated using spectroscopy and microscopy methods, revealing a high presence of stable broken oxygen bonds. Longitudinal traces are then generated by replicating static irradiation structures where the nanoscale modulation can cover partially or completely the photoinscribed traces. The resulting birefringence, the observed anisotropic light scattering properties, and the capacity to write and erase modulated patterns can be used in designing bulk polarization sensitive devices. Various laser-induced structures with optical properties combining guiding, scattering, and polarization sensitivity are reported. The attached polarization functions were evaluated as a function of the fill factor of the nanostructured domains. The polarization sensitivity allows particular light propagation and confinement properties in three dimensional structures.

Experimental investigation of the 100 keV X-ray dose response of the high-temperature thermoluminescence in LiF:Mg,Ti (TLD-100): theoretical interpretation using the unified interaction model

The dose response of LiF:Mg,Ti (TLD-100) chips was measured from 1 to 50,000 Gy using 100 keV X rays at the European Synchroton Radiation Facility. Glow curves were deconvoluted into component glow peaks using a computerised glow curve deconvolution (CGCD) code based on first-order kinetics. The normalised dose response, f(D), of glow peaks 4 and 5 and 5b (the major components of composite peak 5), as well as peaks 7 and 8 (two of the major components of the high-temperature thermoluminescence (HTTL) at high levels of dose) was separately determined and theoretically interpreted using the unified interaction model (UNIM). The UNIM is a nine-parameter model encompassing both the irradiation/absorption stage and the thermally induced relaxation/recombination stage with an admixture of both localised and delocalised recombination mechanisms. The effects of radiation damage are included in the present modelling via the exponential removal of luminescent centres (LCs) at high dose levels. The main features of the experimentally measured dose response are: (i) increase in f(D)(max) with glow peak temperature, (ii) increase in D(max) (the dose level at which f(D)(max) occurs) with increasing glow peak temperature, and (iii) decreased effects of radiation damage with increasing glow peak temperature. The UNIM interpretation of this behaviour requires both strongly decreasing values of ks (the relative contribution of localised recombination) as a function of glow peak temperature and, as well, significantly different values of the dose-filling constants of the trapping centre (TC) and LC for peaks 7 and 8 than those used for peaks 4 and 5. This suggests that different TC/LC configurations are responsible for HTTL. The relative intensity of peak 5a (a low-temperature satellite of peak 5 arising from localised recombination) was found to significantly increase at higher dose levels due to preferential electron and hole population of the trapping/recombination complex giving rise to composite glow peak 5. It is also demonstrated that possible changes in the trapping cross section of the LC and the competitive centres due to increasing sample/glow peak temperature do not significantly influence these observations/conclusions.