Physics in medicine and biology最新文献

Objective. Magnetic resonance imaging is essential in clinical practice due to its non-invasive nature and superior soft-tissue contrast. However, long acquisition times remain a major limitation, leading to motion artifacts and patient discomfort. Most deep learning approaches rely on fully sampled datasets, which are often difficult to obtain. This study aims to develop a self-supervised framework for MRI reconstruction that eliminates dependence on fully sampled training data.Approach. We propose consistency self-supervised learning with ISRVN (CSSL-ISRVN), a novel consistency self-supervised reconstruction framework. At its core lies SRVN, which combines an improved variational network (IVN) with a sensitivity refinement module (SRM). The IVN integrates a feature refinement and denoising module (FRDM), composed of residual blocks (RB) and a Gaussian context transformer (GCT), to jointly extract local and global features. Meanwhile, SRM iteratively refines a task-driven implicit sensitivity modulation variable using the previously reconstructed fullk-space in a reconstruction-sensitivity closed loop, adaptively modulating the multi-coil forward model and data consistency to reduce error accumulation induced by fixed auto-calibration signals-based sensitivity estimates under undersampling. Building on ISTANet and SRVN, we develop ISRVN, a heterogeneous alternating cascade of ISTANet and SRVN: ISTANet first suppresses undersampling artifacts in the multi-coil complex domain to provide cleaner intermediates, enabling SRVN to perform encoding-modulated, physics-consistent refinement for improved reconstruction quality and stability. To eliminate dependence on fully sampled data, we introduce a consistency self-supervised scheme that re-undersamples the originalk-space to train two pairs of consistency networks using calibration and consistency losses.Results. Experiments on three public datasets show that CSSL-ISRVN consistently surpasses existing self-supervised and scan-specific methods, particularly under 1D undersampling masks. It achieves performance competitive with state-of-the-art supervised models.Significance. CSSL-ISRVN offers an effective solution for accelerated MRI reconstruction without fully sampled labels. Its integration of sensitivity refinement, hybrid modeling, and consistency self-supervision enables robust, high-fidelity reconstructions, underscoring its potential for real-world clinical deployment.

Objective: Besides preferential tumour targeting, particle therapy has an increased relative biological effectiveness compared to X-rays, but uncertainties in this effectiveness prevent full exploitation of its clinical benefits. Mechanistic radiation response models can predict the effects of different radiation qualities but the model detail required to capture experimental data remains unclear. In this work, key DNA damage and chromosomal aberration endpoints were simulated and compared to experimental literature. Approach: The TOPAS-nBio and Medras models were used to simulate DNA double strand break (DSB) damage for different radiation exposures. The repair of this damage was simulated, modelling misrepair and chromosome aberration formation in an updated Medras DNA repair model. The characteristic rejoining range of DSB ends in the repair model was re-optimised against experimental photon dose data and tested against ion exposures. Main results: For DSBs, predictions were higher than experimental observations, attributed to the assay resolution limits. Predicted photon-induced chromosome aberrations were higher than observed, with a Root Mean Square Deviation (RMSD) of 1.28 and 1.41 for the Medras and TOPAS-nBio models respectively against the experimental data. The RMSD against the experimental data was lowered by over 70% for both models by re-optimisation of the analytically predicted characteristic DSB end rejoining range to a value of 0.0335 +- 0.0034 (80% of the previous value). This optimisation also performed well when predicting the dependence on ion LET, reducing the proton RMSD by 40% to 0.43 and 0.69 for the Medras and TOPAS-nBio models respectively. Significance: The Medras biological response model was updated and predicted good agreement in aberration yields with the experimental data for both the detailed TOPAS-nBio and less detailed Medras damage models. This highlights how simple mechanistic models, with the guidance of robust experimental data, can be used to explore the effects of radiation quality and guide future experiments.

Objective.Microcalcification (µCalc) detection plays an important role in breast cancer screening. Electronic noise in energy-integrating detectors (EIDs) is the major challenge for this task in current breast cone-beam CT (bCBCT) due to the tight dose constraint for breast imaging. bCBCT with a photon counting detector (PCD) can potentially offer a higher spatial resolution and lower noise. This study performed a direct comparison of bCBCTs with the two detector types via GPU-based Monte Carlo (MC) simulation.Approach.We employed Virtual Clinical Trial for Regulatory Evaluation toolkit to generate a realistic breast phantom with a 0.25mm3voxel size, 80% fat fraction and 14 cm diameter. We considered a bCBCT system with a 60 kV x-ray source filtered with 0.3 mm Cu and detector response functions for PCD and EID. A total of 360 projections were simulated with a total number of3.15×1012photons, corresponding to ∼4 mGy mean glandular dose, comparable to a two-view mammography. We modified our GPU-based MC simulation code to incorporate analytical descriptions ofµCalcs of spherical shapes with diameters ranging from 0.1 to 0.4 mm, in 0.1 mm increments, into the voxelized phantom. A nichrome wire with 0.07 mm diameter was simulated to calculate the modulation transfer functions (MTFs). bCBCT images were reconstructed with the Feldkamp-Davis-Kress algorithm, and image quality andµCalc detection performance were evaluated.Main results.EID-bCBCT had more profound image noise due to electronic noise. The image intensity standard deviations estimated within a region of interest were 0.055 cm^-1for EID-bCBCT and 0.038 cm^-1for PCD-bCBCT, respectively.µCalcs and breast anatomy such as ligaments were more visible in the PCD-bCBCT images. The 10% MTF cutoffs were 5.5 and 9.5 lp mm^-1for EID-bCBCT and PCD-bCBCT, respectively. Contrast-to-noise ratio ranged in 1.20-9.13 for EID-bCBCT and 3.07-14.74 for PCD-bCBCT, depending onµCalc sizes.Significance.We compared EID- and PCD-based bCBCT forµCalc detection using GPU-based MC simulations in a clinically realistic setting. Our results demonstrate a potential advantage of PCD-bCBCT for this detection task.

Objective: HDR brachytherapy is a widely adopted modality for cancer treatment. However, it is not free from error and uncertainty. In-vivo dosimetry (IVD) is the only technique that confirms correct dose delivery. This study details and validates a calibration method for Plastic Scintillation Detector (PSD), bypassing dose gradient and positioning issues in brachytherapy calibration.

Approach: The PRB-0057 PSD (Medscint, Québec, Canada), a 1x1-mm scintillating fiber coupled to a 20-m Eska GH-4001 clear optical fiber (Mitsubishi Rayon, Japan) of the same diameter, was calibrated and connected to the Hyperscint-RP200 research platform for optical signal collection. Hyperspectral calibration was performed at a LINAC with a 6-MV beam, enabling the independent stem effect removal. For validation, brachytherapy measurements with a Sk=29447-U Iridium-192 Flexisource (Elekta Brachy, The Netherlands) were performed in a motorized IBA-Blue-Phantom2 water tank. Dose rates were measured at 10Hz along the source's vertical z-axis at a fixed transverse distance of 1.2±0.05-cm in 0.2-cm steps. Calibration accuracy was evaluated using relative differences (RD) between measured and TG-43U1 dose rates, converted to inverse-square equivalent positional errors. A detailed uncertainty budget was established to the measurement setup.

Main results: Comparison agreed with RDs to around 2.5% at 1.2-cm, corresponding to positional uncertainties of <0.15-mm. At greater distances, up to 8-cm, RDs increase to about 5%, corresponding to positional uncertainties up to 3-mm, mainly due to reduced light-yield. Uncertainties found to depend on the source-detector distance, ranging from 3.92% to 6.42% (k=1) over the range of explored distances.

Significance: Results confirm the effectiveness of a 6-MV external beam PSD calibration to be used in time-resolved IVD. Uncertainties close to the source are consistent with the Afterloader/IBA motorized unit reproducibility and are mainly dominated by reduced detector sensitivity at larger distances. Our study further underlined the intrinsic limitation of IVD in the face of known uncertainties.

Objective: AAPM Task Group Report 203 (TG203) recommends implantable cardiac pacemaker (ICP) dependent patients receive elevated precautions and that maximum dose rates to ICPs should be less than 0.2 Gy/min. Unfortunately, there is little quantitative data published about how ICPs are impacted by radiation dose rates near 0.2 Gy/min. The objective of this paper is to report on modern ICP behavior in response to dose rates near this suggested limit.

Approach: Five single chamber (Azure XT SR) and five dual chamber (Azure XT DR) ICPs were submerged in 0.45% saline solution and placed 1 cm outside the field edge of a 10x10 cm 2 beam. Photon beams of energies 6 MV, 6 FFF, 10 MV, and 10 FFF were tested, with dose rates ranging from 0.15 to 0.92 Gy/min on Varian True Beam and Ethos LINACs. Real-time electrogram data were collected and analyzed for any malfunctions while the beam was on and off. Battery life of each ICP was also checked before irradiation and 6 months post-irradiation.

Main results: No ICP malfunctions or artifacts were observed, even as dose rate increased. Battery life showed normally expected depletion 6 months post irradiation.

Significance: This work studies ICP functionality in clinically relevant radiation therapy conditions and measures the performance of the devices in the context of an ICP dependent patient.

Objective: This work studies the dependency of the charge collection efficiency (CCE) on the atmospheric pressure of commercial ionization chambers (ICs) using an ultra-high dose-per-pulse (DPP) electron beam and proves a theoretical model of this phenomenon. Approach: A custom-made PMMA water phantom, water- and pressure-tight sealed, was connected to a vacuum/pressure pump to vary its inside pressure from 900 hPa to 1100 hPa. ICs were placed at different depths in water and irradiated with ultra-high DPP electron beams (20 MeV, 1.2 μs and 2.0 μs pulse duration). Chamber signals were air-density- and polarity-corrected as a function of DPP (0.1 Gy - 6.5 Gy) at different pressures. The actual DPP was determined and monitored using a calibrated PTW flashDiamond and a current transformer, respectively. Experimental CCEs were compared with numerical calculations, describing the charge transport inside ICs, including free electrons and electric field distortion. Main results: The CCE decreased with increasing pressure (-6%/100 hPa and -8%/100 hPa for the Advanced Markus, and Roos IC, respectively) and followed "logistic functions" with DPP. The CCE0 at a given pressure P0 can be obtained from a CCE1 at different pressure P1 (scaling rule) as: CCE0(P0,DPP) = CCE1(P1,DPP*(P0/P1)^2). Our theoretical model predicted the relative variation of the CCE with pressure, with residuals <3%. The effect can be corrected using the scaling rule even at small changes in pressure (~35 hPa), which can cause a 2% deviation on the CCE, without adding significant CCE uncertainty (~0.1%) for a DPP up to 6.5 Gy per pulse. Significance: This work proposes a scaling rule to correct for recombination losses dependent on atmospheric pressure in ultra-high DPP electron beams. The proposed scaling rule provides a simple, low-uncertainty correction that can be applied to empirically predetermined CCE functions, improving the accuracy of commercial IC-based dosimetry in ultra-high DPP electron beams. .

Objective: Residual gradient non-linearity (GNL) distortions persist after vendor-supplied correction. This work describes and evaluates the performance of a method entitled Phantom-based Residual Error Correction using Individualized System Estimates (PRECISE) to correct these residual distortions. Approach: PRECISE correction maps were created using a commercial distortion phantom on a 0.5T head-specific MR system and added to existing gradient unwarping. The performance of the PRECISE method was then evaluated using the commercial distortion phantom acquired 3D gradient recalled echo acquisitions along each cardinal axis and various other clinical protocols. Residual distortions were decomposed into main magnetic field and gradient non-linearity (GNL) induced components. Main Results: The PRECISE method removed the GNL component of distortions with 95th percentile GNL distortions being reduced from 0.36mm to 0.07mm. Average distortions across various clinical protocols were reduced from 0.193 ± 0.003 mm to mean = 0.055 ± 0.004 mm. 95th percentile distortions across various clinical protocols were reduced from 0.34 ± 0.01 mm to 0.11 ± 0.01 mm. Significance: The phantom-based distortion method in the absence of significant B0 induced distortions results in dramatically improved geometric fidelity. .

Objective: To systematically assess the accuracy and computational performance of a newly proposed stochastic differential equation (SDE)-based model for proton beam dose calculation by benchmarking it against Geant4 in a set of simplified but increasingly challenging phantom geometries. Approach: Building on previous work in Crossley et al. (2025), where energy deposition from a proton beam was modelled using an SDE framework, we implemented the model using standard approximations to interaction cross sections and mean excitation energies, enabling straightforward adaptation to new materials and configurations. The model was benchmarked against Geant4 in homogeneous, longitudinally heterogeneous and laterally heterogeneous phantoms, for assessment of depth-dose behaviour, lateral transport and impact of material heterogeneities. Main results: Across all phantom configurations and beam energies, the SDE model reproduced the main depth-dose characteristics predicted by Geant4, with proton range agreement within 0.2 mm for 100 MeV beams and within 0.6 mm for 150 MeV beams. Voxel-wise comparisons yielded gamma pass rates exceeding 95% for all cases under strict 2%/0.5 mm criteria with a 1% dose threshold. Differences between the two approaches were spatially localised and primarily associated with regions of steep dose gradients or material heterogeneities, while overall lateral beam dispersion was well reproduced. In terms of computational performance, the SDE model achieved speed-up factors of approximately 2.5-3 relative to single-threaded Geant4, consistently across different Geant4 physics lists. Significance: These results demonstrate that the SDE-based approach can reproduce key dosimetric features predicted by high-fidelity Monte Carlo simulations with good accuracy while already offering a moderate reduction in computational cost. Owing to its formulation, the method is naturally amenable to parallel and GPU-accelerated implementations, suggesting potential for substantial further speed improvements. This makes the approach a promising candidate for fast dose calculations in proton therapy.

Objective To investigate Machine Learning (ML) methodologies for predicting glandular dose conversion coefficients for breast models with patient-specific fibroglandular distribution (Γpatient) in digital mammography (DM) and digital breast tomosynthesis (DBT). Approach We investigated four ML algorithms for predicting Γpatient, namely Generalized Additive Model (GAM), XGBoost, Support Vector Regression (SVR) and Automatic Relevance Determination Regression (ARDR). These were trained with Γpatient data generated with a Monte Carlo software and by adopting a dataset of 126 digital breast phantoms with patient-specific fibroglandular distribution. The ML input features were the compressed breast thickness, the glandular fraction by volume and the total breast volume. DM was simulated at 28 kV (Anode/Filter: W/Rh) and at 36 kV (Anode/Filter: W/Al); DBT at 28 kV (Anode/Filter: W/Rh) and 50 degrees scanning angle. Results The four investigated algorithms predicted the Γpatient coefficients with an average difference from the ground truth between -2% (SVR) and +7% (XGBoost). The best model from the GAM fine tuning required the sole compressed breast thickness as input feature. This algorithm presented the smallest model uncertainty, and the lowest cases of dose underestimate. Conclusions The GAM algorithm predicted Γpatient with an average difference from the expected value of 4%, in line with the other investigated algorithms. This algorithm showed the best performance in terms of model uncertainty, with average total estimated uncertainty of 12%, including the model accuracy, for DM at 28 kV. No relevant differences were observed in the case of DBT; bias and uncertainty of the prediction reduced for higher tube voltages. .

Objective: Ionizing radiation is a major source of biological hazard, bearing a broad range of detrimental lesions of the DNA structural and molecular integrity - often leading to genomic instabilities and severe cellular outcomes. The sensitivity of the DNA molecule to ionizing radiation is deeply affected by a variety of chemical and biophysical factors controlling its dynamical behavior, as well as by diverse cellular processes and response pathways. With the available literature offering no clear-cut, conclusive perspective, this work aims at characterizing the role of supercoiling in modulating the mechanical response of DNA to double strand breaks, and discuss the outcome within the broader framework of DNA radiosensitivity. Approach: We assess the linearization of a supercoiled 672-bp DNA minicircle by double strand breaks, i.e., the disruption of the covalent DNA backbone on both complementary strands of the double helix, and a fingerprint lesion of ionizing radiation. To this effect, we employ classical coarse-grained molecular dynamics simulations, and verify how the sequence and supercoiling regime of the minicircle affect the kinetics of the rupturing process. Main results: We observe that the excess torsional stress overall enhances the likelihood of the DNA rupturing but in one specific scenario - associated with a biologically-significant level of negative superhelical density - thereby highlighting a strongly non-symmetric behavior between positive and negative supercoiling regimes. Significance: This work deals critical dynamical insights on the role of topology in the mechanical response of DNA to double strand breaks: Together with the decrease of the effective volume of a DNA target enforced by an excess/defect of superhelical density, we infer that a degree of supercoiling belonging to an average biological scenario might factor in the earliest radiobiological response of (naked) DNA.