Physics in medicine and biology最新文献

Objective.This study aimed to determine the optimal coil orientation for transcranial magnetic stimulation (TMS) of the hand motor area by integrating physiological and computational approaches.Approach.Resting motor thresholds (RMTs) were measured in 10 healthy volunteers for the first dorsal interosseous (FDI) and abductor digiti minimi (ADM) muscles when stimulating the primary motor cortex (M1) with a coil set at several orientations ranging from 0° to 90°. Electric field (EF) distributions were estimated using individualized head models constructed from magnetic resonance imaging (MRI) data of the same 10 participants in the measurements, as well as additional 135 MRI-derived models. Simulations employed a scalar-potential finite-difference method to quantify the EF magnitude in the M1-hand region across orientations.Main results.The lowest RMTs were obtained between 30° and 45° for both muscles, and the optimal angle depended on the target muscle (44.3 ± 9.8° for FDI and 37.9 ± 10.3° for ADM). Calculated EF magnitudes correlated negatively with the measured RMT values. Analysis of the additional 135 MRI-derived head models showed that coil orientations of 30°-45° most frequently produced high EF, and this coil-angle dependence remained stable even when the coil position was slightly displaced. Anatomical analysis revealed that the cortical surface orientations in the M1-hand area were frequently 30°-45° to the parasagittal plane.Significance.These findings support the current guidelines' recommendation of a ∼45° orientation, but suggest that a 30°-45° range better aligns EF with cortical geometry. Individualized optimization can further improve the precision and efficacy of TMS.

Objective: To evaluate whether clinical megavoltage radiotherapy beams can function as dual-purpose sources that deliver therapeutic dose while simultaneously generating high-flux entangled 511 keV photon pairs for quantum ghost imaging and Quantum Theranostics (QTX). Approach. Geant4 Monte Carlo simulations modeled water-equivalent spherical phantoms containing gold nanoparticle (10 mg/mL)-loaded tumors irradiated with 6, 10, and 15 MV clinical beams. We quantified entangled photon-pair yields, positronium lifetime sensitivity, Doppler and Compton broadening, voxel-level signal-to-noise ratios (SNR), and entanglement retention as functions of depth and beam energy, incorporating detector performance and coincidence gating. Main results. AuNP-loaded tumors produced entangled photon-pair yields of 5.0×10⁷-1.5×10⁸ pairs Gy⁻¹ cm⁻³ (10⁷-10⁹ pairs s⁻¹ at clinical dose rates), with per-voxel SNR of 128-44,395 across 27 configurations (5-15 cm depth, 5-20 mm tumor radius). Doppler broadening (≈1-3 keV), AuNP-induced line broadening (≈0.5-1 keV), and Compton shifts (≈13-171 keV) provided spectroscopic sensitivity to tissue composition, nanoparticle uptake, and microenvironmental heterogeneity, while depth-dependent coherence analysis showed that a substantial fraction of entangled pairs survive to support ghost imaging at clinical depths.

Significance: These results indicate that clinical MV beams can act as practical high-flux entangled photon sources, enabling simultaneous therapy and quantum-enhanced imaging. By combining time-resolved positronium lifetimes with energy-resolved Doppler and Compton spectroscopy, the proposed QTX platform could deliver real-time mapping of tumor microenvironment and composition during treatment, extending quantum imaging concepts from optical to therapeutic energy scales. .

Diffusion magnetic resonance imaging is a non-invasive technique used to characterize tissue microstructure by measuring the diffusion of water molecules. Conventional Q-space trajectory imaging (QTI) estimates diffusion using low-order moments; however, it often neglects higher-order moments, such as the skewness tensor, resulting in an incomplete representation of diffusion asymmetry and potential estimation bias. In this work, we propose QTI with skewness tensor constraints, a method that incorporates higher-order skewness tensors under positivity constraints to mitigate deviations in the estimation of lower-order moments caused by the omission of higher-order asymmetry information. Furthermore, we introduce linear trace-weighted and quadratic trace-weighted filters to enhance high-diffusion components while suppressing low-diffusion components. Extensive experiments conducted on public, noisy, and synthetic datasets demonstrate that our method yields estimates closer to the ground truth on synthetic data and exhibits superior robustness in noisy conditions.

Objective. Multimodal images contain complementary information that is valuable for deep learning (DL)-based image segmentation. To enable effective multimodal feature learning and fusion for accurate segmentation, multimodal images usually need to be registered to achieve anatomical alignment. However, in clinical settings, multimodal image registration is often challenging. For instance, to reduce radiation exposure, CT scans usually have a smaller field of view than MR, i.e. inconsistent anatomical content in CT and MR images, hindering accurate registration. Using such misaligned multimodal images, segmentation performance could be significantly degraded. This study aims to develop a DL-based multimodal image segmentation method that is capable of learning high-quality and strongly related image features from misaligned multimodal images without registration and produce accurate segmentation results comparable to that obtained with well-aligned multimodal images.Approach. In our method, a unified body space (UBS) module is presented, where image patches cropped from misaligned modalities are encoded to positions and projected into a UBS, thereby largely mitigating the misalignment among multimodal images. Built upon the UBS module, a new spatial-attention is proposed and integrated into a multilevel feature fusion (MFF) module, where features learned from misaligned multimodal images are effectively fused at internal-, spatial-, and modal-levels, leading the segmentation of misaligned multimodal images to a high accuracy level.Main results. We validate our method on both public and in-house multimodal image datasets containing 1472 patients. Experimental results demonstrate that our method outperforms state-of-the-art methods. The ablation study further confirms that the UBS modules can accurately project image patches from different modalities into the UBS. Moreover, the internal-, spatial-, and modal-level feature fusion in the MFF module substantially enhances segmentation accuracy for misaligned multimodal images.Significance. Our method presents a new registration-free multimodal segmentation framework that explicitly models the correspondence between image patches and anatomical positions, enabling effective fusion of misaligned modalities and improved segmentation performance in realistic clinical scenarios. Codes of our method are available athttps://github.com/BH-MICom/Patch2Space.

Objective. Evaluate the dosimetric performance of a novel three-dimensional (3D) electronic detector array for radiotherapy quality assurance using Monte Carlo simulations.Approach. Monte Carlo simulations were performed with Geant4 and MCNP6.2. A detailed detector model (MC_Model) was implemented consisting of a 50 × 50 × 50 cm³polymethyl methacrylate (PMMA) phantom with 20 imbedded active matrices (AMs) at strategic depths, 1169 pixels per AM, and 23 380 pixels for the entire detector. A pixel comprises a diode, capacitor, and MOSFET, where the diode elements provide a 42 × 42 cm²sensitive area within an AM. Photon beam energy spectra of 6 MV and 10 MV, respectively, of a Varian radiotherapy linear accelerator (linac) were used in the simulations. Dosimetric data consisting of per cent depth-doses (PDDs) and cross-/in-plane profiles for field sizes of 5 × 5 cm²through 40 × 40 cm²were simulated for the MC_Model, and a homogeneous PMMA phantom (MC_PMMA) of similar dimensions.Main results.MC_Modelversus MC_PMMAPDD data difference for Geant4 and MCNP6 were within 3.44% and 3.83%, while profiles (cross-/in-plane) were within 5.54% and 5.68%, for all field sizes and energies.Significance. These results suggest that a 3D electronic detector could provide suitable dosimetric data for radiotherapy QA, and if realised, it could likely provide it in less time than current methods.

Objective.Rapid and accurate mapping of brain tissue pH is crucial for early diagnosis and management of ischemic stroke. Amide proton transfer (APT) imaging has been used for this purpose but suffers from hypointense contrast and low signal intensity in lesions. Guanidine chemical exchange saturation transfer (CEST) imaging provides hyperintense contrast and higher signal intensity in lesions at appropriate saturation power, making it a promising complementary approach. However, quantifying the guanidine CEST effect remains challenging due to its proximity to water resonance and the influence of multiple confounding effects. This study presents a machine learning (ML) framework to improve the accuracy and robustness of guanidine CEST quantification with reduced scan time.Approach.The model was trained onpartially synthetic data, where measured line-shape information from experiments were incorporated into a simulation framework along with other CEST pools whose solute fraction (f_s), exchange rate (k_sw), and relaxation parameters were systematically varied. Gradient-based feature selection was used to identify the most informative frequency offsets to reduce the number of acquisition points.Main results.The proposed model achieved significantly higher accuracy than polynomial fitting, multi-pool Lorentzian fitting, and ML models trained solely on synthetic orin vivodata. Gradient-based feature selection identified the most informative frequency offsets, reducing acquisition points from 69 to 19, a 72% reduction in CEST scan time without loss of accuracy.In vivo, conventional fitting methods produced unclear lesion contrast, whereas our model predicted clear hyperintense lesion maps. The strong negative correlation between guanidine and APT effects supports its physiological relevance to tissue acidosis.Significance.The use of partially synthetic training data combines realistic spectral features with known ground-truth values, overcoming limitations of purely synthetic or limitedin vivodatasets. Leveraging this data with ML, enables robust quantification of guanidine CEST effects, showing potential for rapid pH-sensitive imaging.

Objective. Tumor treating fields (TTFields) is an emerging cancer therapy whose efficacy is closely linked to the electric field (EF) intensity delivered to the tumor. However, current computational workflows for simulating the EF and planning treatment rely on time-consuming manual segmentation and proprietary software, hindering efficiency, reproducibility, and accessibility.Approach. We introduce AutoSimTTF, a fully automatic pipeline for personalized EF simulation and optimized treatment planning for TTFields. The end-to-end workflow utilizes advanced deep learning model for automated tumor segmentation, conducts finite element method-based EF simulation, and determines a computationally optimized treatment plan via a novel, physics-based parameter optimization method.Main results. The automated segmentation module achieved high precision, yielding a Dice similarity coefficient of 0.91 for the whole tumor. In terms of efficiency, the active planning workflow was completed in approximately 12 min, significantly outperforming conventional multi-day manual processes. The pipeline's simulation accuracy was validated against a conventional semi-automated workflow, demonstrating deviations of less than 14.1% for most tissues. Critically, the parameter optimization generated personalized transducer montages that produced a significantly higher EF intensity at the tumor site (up to 111.9% higher) and substantially improved field focality (19.4% improvement) compared to traditional fixed-array configurations.Significance. AutoSimTTF addresses major challenges in efficiency and reproducibility, paving the way for data-driven personalized TTFields therapy and large-scale computational research.

Objective.Geometric distortions in diffusion-weighted imaging (DWI) compromise accurate tumor delineation and spatial localization, limiting its utility in radiation therapy planning and response monitoring. These distortions can be corrected through multimodal registration between distorted DWI and undistorted anatomical images, while conventional mutual information-based optimization often fails due to local minima and produces non-smooth, physically implausible deformations.Approach.This study proposes a landmark matching B-spline implicit neural representation framework for DWI distortion correction. The method integrates anatomical correspondences from a foundation landmark matching model with B-spline parameterized deformation fields to overcome local minima inherent in mutual information optimization. The framework employs Fourier-encoded multi-layer perceptrons to model B-spline deformation fields while ensuring physically plausible transformations, enabling robust multimodal registration between distorted DWI and anatomical references.Main results.Evaluation on brain and abdominal datasets demonstrated superior performance compared to established methods. The proposed approach achieved average Dice coefficients of 0.919 ± 0.038 (brain) and 0.926 ± 0.032 (abdomen), significantly outperforming all baseline methods. On simulated data, our method achieved an average PSNR of 25.912 ± 3.148 dB, NCC of 0.911 ± 0.137, and SSIM of 0.888 ± 0.107, the best among all methods.Significance.By combining the regularization properties of B-spline parameterization with the cross-modal matching capabilities of foundation models, our method achieves more accurate correction of geometric distortions in DWI, with the potential to enhance the precision of intra/post-radiotherapy assessment.

Objective.Despite major advances in dual-energy computed tomography (CT), obtaining accurate attenuation values for quantitative applications remains a technical challenge. To address this topic, we introduce a novel projection data domain material decomposition method that is an extension of an approach we recently proposed for beam hardening correction in single energy CT.Approach.The proposed method employs object-specific scatter correction and an analytical energy response model. We compare its performance to image-based material decomposition on accuracy of attenuation values using the American College of Radiology (ACR) CT accreditation phantom, scanned with consecutive low and high energy axial scans in centered and off-centered positions. Accuracy is assessed across the five inserts, and the images are analyzed for beam hardening artifacts and noise. Additionally, we assess the usefulness of object-specific scatter correction, and we assess performance over conventional data domain material decomposition and for anthropomorphic abdomen phantom imaging.Main results.In the ACR phantom, the proposed method yielded a significant improvement in accuracy of the attenuation values, particularly at low energies (<70 keV), and an important reduction in beam hardening artifacts. While similarly high accuracy was achieved for water, quantitative error within the non-water inserts was lower and more uniform across the 30-140 keV range, especially in the more challenging off-centered positioning of the phantom. Noise showed expected parabolic behavior, but with minimum at lower keV, which may be clinically advantageous. Object-specific scatter correction was shown to prevent major artifacts. Advantages over conventional data-domain decomposition clearly appeared when only a standard phantom is available to calibrate the latter. Lastly, the proposed method was shown to perform well, without any changes, in the more complex scenario of abdominal phantom imaging.Significance.This work demonstrates that data-based material decomposition using an analytical energy response model with object-specific scatter correction offers a promising pathway to improve accuracy of CT attenuation values.

Saturation transfer MR fingerprinting (ST-MRF) is a quantitative molecular MRI method that simultaneously estimates parameters of free water, solute, and semisolid macromolecule protons. The accuracy of these quantification is highly dependent on the choice of acquisition parameters, and thus, the optimization of the data acquisition schedule is crucial to improve acquisition efficiency and quantification accuracy. Herein, we developed a learning-based optimization framework for ST-MRF, incorporating a deep Bloch equation simulator as a surrogate model for the forward Bloch equation solver to enable rapid simulations. Notably, the deep Bloch equation simulator overcomes the non-differentiability of the original model by enabling gradient computation during backpropagation within the physics-informed optimization framework, thereby allowing iterative updates of the acquisition schedule to minimize quantification error. In addition, the proposed method estimated an accurate ΔB₀map with the inclusion of a minimal number of scans to address B₀inhomogeneity. B₁inhomogeneity was corrected by providing a relativeB₁map as an input to the quantification network. We validated our approach using Bloch-McConnell equation-based digital phantoms and further evaluated the performance of the proposed optimized ST-MRF framework inin vivoexperiments. Our results showed that the optimal ST-MRF schedule outperformed other data acquisition schedules with regard to quantification accuracy. In addition, we enhanced thein vivoquantitative maps by correcting motion artifacts and suppressing noise using self-supervised learning techniques. The optimal ST-MRF approach could generate accurate and reliable multi-tissue parameter maps within a clinically acceptable time.