Physics in medicine and biology最新文献

Objective.Among various deep-network-based sparse-view CT image reconstruction studies, the sinogram upscaling network has been predominantly employed to synthesize additional view information. However, the performance of the sinogram-based network is limited in terms of removing aliasing streak artifacts and recovering low-contrast small structures. In this study, we used a view-by-view back-projection (VVBP) tensor-domain network to overcome such limitations of the sinogram-based approaches.Approach.The proposed method offers advantages of addressing the aliasing artifacts directly in the 3D tensor domain over the 2D sinogram. In the tensor-domain network, the multi-planal anti-aliasing modules were used to remove artifacts within the coronal and sagittal tensor planes. In addition, the data-fidelity-based refinement module was also implemented to successively process output images of the tensor network to recover image sharpness and textures.Main result.The proposed method showed outperformance in terms of removing aliasing artifacts and recovering low-contrast details compared to other state-of-the-art sinogram-based networks. The performance was validated for both numerical and clinical projection data in a circular fan-beam CT configuration.Significance.We observed that view-by-view aliasing artifacts in sparse-view CT exhibit distinct patterns within the tensor planes, making them effectively removable in high-dimensional representations. Additionally, we demonstrated that the co-domain characteristics of tensor space processing offer higher generalization performance for aliasing artifact removal compared to conventional sinogram-domain processing.

Objective: To develop an empirical model to predict carbon ion (C-ion) relative biological effectiveness (RBE). Approach. We used published cell survival data comprising 360 cell line/energy combinations to characterize the linear energy transfer (LET) dependence of cell radiosensitivity parameters describing the dose required to achieve a given survival level, e.g. 5% (D_5%), which are linearly correlated between photon and C-ion radiations. Based on the LET response of the metrics D5% and D37%, we constructed a model containing four free parameters that predicts cells' linear quadratic model (LQM) survival curve parameters for C-ions, α_Cand β_C, from the reference LQM parameters for photons, α_Xand β_X, for a given C-ion LET value. We fit our model's free parameters to the training dataset and assessed its accuracy via leave-one out cross-validation. We further compared our model to the local effect model (LEM) and the microdosimetric kinetic model (MKM) by comparing its predictions against published predictions made with those models for clinically relevant LET values in the range of 23-107 keV/μm. Main Results. Our model predicted C-ion RBE within ±7%-15% depending on cell line and dose which was comparable to LEM and MKM for the same conditions. Significance. Our model offers comparable accuracy to the LEM or MKM but requires fewer input parameters and is less computationally expensive and whose implementation is so simple we provide it coded into a spreadsheet. Thus, our model can serve as a pragmatic alternative to these mechanistic models in cases where cell-specific input parameters cannot be obtained, the models cannot be implemented, or for which their computational efficiency is paramount.

Purpose.Cherenkov imaging during radiotherapy provides a real time visualization of beam delivery on patient tissue, which can be used dynamically for incident detection or to review a summary of the delivered surface signal for treatment verification. Very few photons form the images, and one limitation is that the noise level per frame can be quite high, and mottle in the cumulative processed images can cause mild overall noise. This work focused on removing or suppressing noise via image postprocessing.Approach.Images were analyzed for peak-signal-to-noise and spatial frequencies present, and several established noise/mottle reduction algorithms were chosen based upon these observations. These included total variation minimization (TV-L1), non-local means filter (NLM), block-matching 3D (BM3D), alpha (adaptive) trimmed mean (ATM), and bilateral filtering. Each were applied to images acquired using a BeamSite camera (DoseOptics) imaged signal from 6x photons from a TrueBeam linac delivering dose at 600 MU min^-1incident on an anthropomorphic phantom and tissue slab phantom in various configurations and beam angles. The standard denoised images were tested for PSNR, noise power spectrum (NPS) and image sharpness.Results.The average peak-signal-to-noise ratio (PSNR) increase was 17.4% for TV-L1. NLM denoising increased the average PSNR by 19.1%, BM3D processing increased it by12.1% and the bilateral filter increased the average PSNR by 19.0%. Lastly, the ATM filter resulted in the lowest average PSNR increase of 10.9%. Of all of these, the NLM and bilateral filters produced improved edge sharpness with, generally, the lowest NPS curve.Conclusion.For cumulative image Cherenkov data, NLM and the bilateral filter yielded optimal denoising with the TV-L1 algorithm giving comparable results. Single video frame Cherenkov images exhibit much higher noise levels compared to cumulative images. Noise suppression algorithms for these frame rates will likely be a different processing pipeline involving these filters incorporated with machine learning.

Objective.To assess the accuracy and stability of areal bone-mineral-density (aBMD) measurements from multi-energy CT localizer radiographs acquired using photon-counting detector (PCD) CT.Approach.A European Spine Phantom (ESP) with hydroxyapatite (HA 0.5, 1.0 and 1.5 g cm^-2) was scanned using clinical PCD-CT and a dual-energy x-ray absorptiometry (DXA) to compare aBMD values. To assess aBMD stability and reproducibility, PCD-localizers were acquired twice a day for one week, and once per week for five weeks. Multiple ESP anteroposterior thicknesses (18, 26, 34, and 40 cm) were achieved using a synthetic gel layer and scanned across eight tube current values for both 120 kV (15-120 mA) and 140 kV (10-80 mA). One-way analysis of variance was performed for statistical significance (p< 0.05 considered significant). Quantitative HA and water maps were reconstructed using a prototype material-decomposition software, and aBMD was calculated after background correction.In vivoperformance of PCD-based aBMD was illustrated using a patient scan acquired at 140 kV and 40 mA, and lumbar aBMD values were compared with DXA.Main results.The ESP aBMD values from PCD-localizers demonstrated excellent day-to-day stability with a coefficient-of-variation ranging from 0.42% to 0.53%, with mean absolute percentage errors (MAPE) of less than 5% for all three vertebral bodies. The coefficient-of-variation for weekly scans ranged from 0.17% to 0.60%, again with MAPE below 5% for all three vertebral bodies. Across phantom sizes and tube currents, the MAPE values varied ranging from 1.84% to 13.78% for 120 kV, and 1.38%-9.11% for 140 kV. No significant difference was found between different tube currents. For the standard phantom size, DXA showed 11.21% MAPE whereas PCD-CT showed 3.04% MAPE. For the patient scan, deviation between PCD-based aBMD values and those obtained from DXA ranged from 0.07% to 9.82% for different lumbar vertebra.Significance.This study highlights the accuracy and stability of PCD-CT localizers for measuring aBMD. We demonstrated aBMD accuracy across different sizes and showed that higher radiation doses did not inherently increase aBMD accuracy.

Objective. Intensity modulated proton therapy (IMPT) is susceptible to uncertainties in patient setup and proton range. Robust optimization is employed in IMPT treatment planning to ensure sufficient coverage of the clinical target volume (CTV) in predefined scenarios, albeit at a price of increased planning times. We investigated a deep learning (DL) strategy for dose predictions in individual error scenarios in head and neck cancer IMPT treatment planning, enabling direct evaluation of plan robustness. The model is able to differentiate between scenarios by using embeddings of the scenario index.Approach. To accommodate resolution disparities in planning CT-scans and accommodate the setup error scenarios, we introduced scenario-specific isocentric distance maps as inputs to the DL models. For 392 H&N cancer patients, high-quality 9-scenario ground truth (GT) robust plans were generated with wish-list driven fully automated multi-criteria optimization. The scenario index is converted to one-hot-vector that is used to derive the scenarios embeddings through the training of the DL model, aiding the model to predict a scenario specific dose distribution.Main results. The model achieved within 1%-point of agreement with the GT the predictedV95%of the voxelwise minimum dose for CTV Low and CTV High for 96% and 75% respectively of the test patients. Considering all robustness scenarios, median differences were 0.035%-point for CTV HighV95%, 0.11%-point for CTV LowV95%, 0.29 GyE for parotidsDmean, 0.7 GyE for submandibular glandsDmeanand 0.9 GyE for oral cavityDmean. Prediction of full 3D dose distributions for all scenarios took around 14 s.Significance. Predicting individual scenarios for robust proton therapy using DL dose prediction is feasible, enabling direct robustness evaluation of the predicted scenario doses.

Objective. Prompt gamma-ray timing is an emerging technology in the field of particle therapy treatment verification. This system measures the arrival times of gamma rays produced in the patient body and uses the cyclotron radio frequency signal as time reference for the beam micro-bunches. Its translation into clinical practice is currently hindered by observed instabilities in the phase relation between the cyclotron radio frequency and the measured arrival time of prompt gamma rays. To counteract this, two proton bunch monitors are presented, integrated into the prompt gamma-ray timing workflow and evaluated.Approach. The two monitors are (a) a diamond detector placed at the beam energy degrader, and (b) a cyclotron monitor signal measuring the phase difference between dee current and voltage. First, the two proton bunch monitors as well as their mutual correlation were characterized. Then, a prompt gamma-ray timing measurement was performed aiming to quantify the present magnitude of the phase instabilities and to evaluate the ability of the proton bunch monitors to correct for these instabilities.Main results. It was found that the two new monitors showed a very high correlation for intermediate proton energies after the first second of irradiation, and that they were able to reduce fluctuations in the detected phase of prompt gamma rays. Furthermore, the amplitude of the phase instabilities had intrinsically decreased from about 700 ps to below 100 ps due to cyclotron upgrades.Significance. The uncertainty of the prompt gamma-ray timing method for proton treatment verification was reduced. For routine clinical application, challenges remain in accounting for detector load effects, temperature drifts and throughput limitations.

Objective. Current reference dosimetry methods for spatially fractionated radiation therapy (SFRT) assume a negligible beam quality change, perturbation, or volume-averaging correction factor. Therefore, the aim of this work was to investigate the impact of the grid collimators on the dosimetric characteristics of a 6 MV photon beam. A detector-specific correction factor,kQgrid,  Qmsr fgrid,fmsr, was proposed. Several dosimeters were evaluated for their ability to measure both reference dose and grid output factors (GOFs).Approach. A Monte Carlo model of a grid collimator was created to study the change in the depth dose characteristics with the grid collimator. The impact of the collimator on the percent depth dose (PDD), electron contamination, and average photon energy was investigated. ThekQgrid,  Qmsr fgrid,fmsrcorrection factors were calculated for two reference-class micro ion chambers. The reference dose and GOFs were measured with a grid collimator using six ion chambers, two silicon diodes, and a diamond detector.Main results.The PDD in the presence of the grid was observed to be steeper compared to the open field. The average photon energy increased from 1.33 MeV to 1.74 MeV with the presence of the grid collimator. The dose contribution by scattered photons was significantly higher at deeper regions for the open field compared to the grid field. ThekQgrid,  Qmsr fgrid,fmsrcorrection was calculated to be <0.5%. The reference dose for all detectors, except for the CC13 and CC04 chambers, was within 1% of each other. The CC13 under-responded up to 3.2% due to volume-averaging effects. The GOFs calculated for all detectors, except Razor and A16, were within 1% of each other.Significance. The phantom scatter dictates the change in the PDD with the presence of the grid. The micro ion chambers exhibit negligiblekQgrid,  Qmsr fgrid,fmsrcorrection. All detectors, except the CC13 ion chamber, were found to be suitable for SFRT reference dosimetry.

The method combining Monte Carlo (MC) simulation and mesh-type cell models provides a way to accurately assess the cellular dose induced byβ-emitters. Although this approach allows for a specific evaluation of various nuclides and cell type combinations, the associated time cost for obtaining results is relatively high. In this work, we propose a Microdosimetric assessment method for Internal exposure ofβ-emitters based on Mesh-type Cell cluster models (abbreviated as MIMC-β). This approach is applied to evaluate the dose in various types of cells (human bronchial epithelial cells, BEAS-2B; normal human liver cells, L-O2; and normal human small intestine epithelial cells, FHs74Int) exposed toβ-emitters. Furthermore, microdosimetric quantity based on the cell cluster model are employed to estimate the relative biological effectiveness (RBE) ofβ-emitters. The results indicate that this method can accurately and rapidly predict cellular doses caused by different types ofβ-emitters, significantly mitigating the efficiency challenges associated with directly employing MC to estimate the overall dose of the mesh-type cell cluster model. In comparison with results obtained from direct simulations of uniform administration ofβ- sources using PHITS for validation, the cellular cluster overallS-values obtained through MIMC-βshow discrepancies mostly below 5%, with the minimum deviation reaching 1.35%. Small sampling sizes within the cell nucleus led to larger average lineal energies. In comparison to C-14, the differences in cellular cluster average lineal energy for Cs-134, Cs-137, and I-131 are negligible, resulting in close numerical estimations of RBE based on lineal energy. The MIMC-βcan be extended to diverse cell types andβ-emitters. Additionally, the RBE assessment based on the cell cluster model offers valuable insights for predicting radiobiological damage resulting from internal exposure byβ-emitters. This method is expected to find applicability in various realistic scenarios, including radiation protection and radioligand therapy.

A 750 kHz, 360-element ultrasound array has been built for transcranial histotripsy applications. This study aims to evaluate its performance to determine whether this array is adequate for treating a wide range of brain locations through a human skull. Treatment location profiles in 2 excised human skulls were experimentally characterized based on passive cavitation mapping. Full-wave acoustic simulations were performed in 8 human skulls to analyze the ultrasound propagation at shallow targets in skulls with different properties. Results showed that histotripsy successfully generated cavitation from deep to shallow targets within 5 mm from the skull surface in the skull with high SDR and small thickness, whereas in the skull with low SDR and large thickness, the treatment envelope was limited up to 16 mm from the skull surface. Simulation results demonstrated that the treatment envelope was highly dependent on the skull acoustic properties. Pre-focal pressure hotspots were observed in both simulation and experiments when targeting near the skull. For each skull, the acoustic pressure loss increases significantly for shallow targets compared to central targets due to high attenuation, large incident angles, and pre-focal pressure hotspots. Strategies including array design optimization, pose optimization, and amplitude correction, are proposed to broaden the treatment envelope. This study identifies the capabilities and limitations of the 360-element transcranial histotripsy array and suggests strategies for designing the next-generation transcranial histotripsy array to expand the treatment location profile for a future clinical trial.

Objective.Radial cine-MRI allows for sliding window reconstruction at nearly arbitrary frame rate, promising high-speed imaging for intra-fractional motion monitoring in magnetic resonance guided radiotherapy. However, motion within the reconstruction window may determine the location of the reconstructed target to deviate from the true real-time position (target positioning errors), particularly in cases of fast breathing or for anatomical structures affected by the heartbeat. In this work, we present a proof-of-concept study aiming to enhance radial cine-MR imaging by implementing deep-learning-based intra-frame motion compensation techniques.Approach.A novel network (TransSin-UNet) was proposed to continuously estimate the final-position image of the target, corresponding to end of the frame acquisition. Within the radial k-space reconstruction window, the spatial-temporal dependencies among the sinogram representation of the spokes were modeled by a transformer encoder subnetwork, followed by a UNet subnetwork operating in the spatial domain for pixel-level fine-tuning. By simulating motion-dependent radial sampling with (tiny) golden angles, we generated datasets from 25 4D digital anthropomorphic lung cancer phantoms. The network was then trained and extensively evaluated across datasets characterized by varying azimuthal radial profile increments.Main Results.The method required additional 4.8 ms per frame over the conventional approach involving direct image reconstruction with motion-corrupted spokes. TransSin-UNet outperformed architectures relying solely on transformer encoders or UNets across all the comparative evaluations, leading to a noticeable enhancement in image quality and target positioning accuracy. The normalized root mean-squared error decreased by 50% from the initial value of 0.188 on average, whereas the mean Dice similarity coefficient of the gross tumor volume increased from 85.1% to 96.2% in the investigated cases. Furthermore, the final-positions of anatomical structures undergoing substantial intra-frame deformations were precisely derived.Significance.The proposed approach enables an effective intra-frame motion compensation, offering an opportunity to reduce errors in radial cine-MR imaging for real-time motion management.