Physics in medicine and biology最新文献

Objective.Bone scans play an important role in skeletal lesion assessment, but gamma cameras exhibit challenges with low sensitivity and high noise levels. Deep learning (DL) has emerged as a promising solution to enhance image quality without increasing radiation exposure or scan time. However, existing self-supervised denoising methods, such as Noise2Noise (N2N), may introduce deviations from the clinical standard in bone scans. This study proposes an improved self-supervised denoising technique to minimize discrepancies between DL-based denoising and full scan images.Approach.Retrospective analysis of 351 whole-body bone scan data sets was conducted. In this study, we used N2N and Noise2FullCount (N2F) denoising models, along with an interpolated version of N2N (iN2N). Denoising networks were separately trained for each reduced scan time from 5 to 50%, and also trained for mixed training datasets, which include all shortened scans. We performed quantitative analysis and clinical evaluation by nuclear medicine experts.Main results.The denoising networks effectively generated images resembling full scans, with N2F revealing distinctive patterns for different scan times, N2N producing smooth textures with slight blurring, and iN2N closely mirroring full scan patterns. Quantitative analysis showed that denoising improved with longer input times and mixed count training outperformed fixed count training. Traditional denoising methods lagged behind DL-based denoising. N2N demonstrated limitations in long-scan images. Clinical evaluation favored N2N and iN2N in resolution, noise, blurriness, and findings, showcasing their potential for enhanced diagnostic performance in quarter-time scans.Significance.The improved self-supervised denoising technique presented in this study offers a viable solution to enhance bone scan image quality, minimizing deviations from clinical standards. The method's effectiveness was demonstrated quantitatively and clinically, showing promise for quarter-time scans without compromising diagnostic performance. This approach holds potential for improving bone scan interpretations, aiding in more accurate clinical diagnoses.

Objective.Focused ultrasound (FUS) is a promising non-invasive therapeutic approach that can be used to generate thermal and non-thermal bioeffects. Several non-thermal FUS therapies rely on FUS-induced oscillations of microbubbles (MBs), a phenomenon referred to as cavitation. Cavitation monitoring in real time is essential to ensure both the efficacy and the safety of FUS therapies. This study aims to introduce a new magnetic resonance (MR) method for cavitation monitoring during FUS therapies.Approach.By finely synchronizing the FUS pulse with an accelerated turbo spin-echo MR sequence, the cavitation effect could be quantitatively estimated on the acquired images at 1-Hz refresh rate. The proposed method was assessed in vitro in a water bath. A series of FUS pulses were generated on a silicone tube filled with MBs at different acoustic pressures (0.07-2.07 MPa) and pulse durations (20-2000μs). MR images and passive cavitation detection (PCD) signals were simultaneously acquired for each FUS pulse.Main results.Inertial cavitation was found to induce a quantitatively interpretable signal loss on the MR image. The transition from stable to inertial cavitation was identified on MR cavitation maps with high repeatability. These results were found to be in good agreement with PCD measurements in terms of pressure thresholds between stable and inertial cavitation. MR cavitation imaging was shown to be sensitive to short and even ultrashort FUS pulses, from 2 ms down to 20μs. The presented theoretical model suggests that the signal loss in MR cavitation imaging relies on susceptibility changes related to the diameter of the oscillating MBs.Significance.The proposed MR cavitation imaging method can both locate and characterize cavitation activity. It has therefore the potential to improve the efficacy and safety of FUS therapies, particularly for localized drug delivery applications.

Objective.To present a long short-term memory (LSTM) network-based dose calculation method for magnetic resonance (MR)-guided proton therapy.Approach.35 planning computed tomography (CT) images of prostate cancer patients were collected for Monte Carlo (MC) dose calculation under a perpendicular 1.5 T magnetic field. Proton pencil beams (PB) at three energies (150, 175, and 200 MeV) were simulated (7560 PBs at each energy). A 3D relative stopping power cuboid covering the extent of the PB dose was extracted and given as input to the LSTM model, yielding a 3D predicted PB dose. Three single-energy (SE) LSTM models were trained separately on the corresponding 150/175/200 MeV datasets and a multi-energy (ME) LSTM model with an energy embedding layer was trained on either the combined dataset with three energies or a continuous energy (CE) dataset with 1 MeV steps ranging from 125 to 200 MeV. For each model, training and validation involved 25 patients and 10 patients were for testing. Two single field uniform dose prostate treatment plans were optimized and recalculated with MC and the CE model.Results.Test results of all PBs from the three SE models showed a mean gamma passing rate (2%/2 mm, 10% dose cutoff) above 99.9% with an average center-of-mass (COM) discrepancy below 0.4 mm between predicted and simulated trajectories. The ME model showed a mean gamma passing rate exceeding 99.8% and a COM discrepancy of less than 0.5 mm at the three energies. Treatment plan recalculation by the CE model yielded gamma passing rates of 99.6% and 97.9%. The inference time of the models was 9-10 ms per PB.Significance.LSTM models for proton dose calculation in a magnetic field were developed and showed promising accuracy and efficiency for prostate cancer patients.

Objective.The linear energy transfer (LET) of proton therapy beams increases rapidly from the Bragg peak to the end of the beam. Although the LET can be determined using analytical or computational methods, a technique for efficiently measuring its spatial distribution has not yet been established. Thus, the purpose of this study is to develop a technique to measure the two-dimensional LET distribution in proton therapy in real time using a combination of multiple scintillators with different quenching.Approach.Inorganic and organic scintillator sheets were layered and irradiated with proton beams. Two-color signals of the CMOS sensor were obtained from the scintillation light and calibration curves were generated using LET. LET was calculated using Monte Carlo simulations asLETtandLETdweighted by fluence and dose, respectively. The accuracy of the calibration curve was evaluated by comparing the calculated and measured LET values for the 200 MeV monoenergetic and spread-out Bragg peak (SOBP) beams. LET distributions were obtained from the calibration curves.Main results.The deviation between the calculated and measured LET values was evaluated. For bothLETtandLETd, the deviation in the plateau region of the monoenergetic and SOBP beams tended to be larger than those in the peak region. The deviation was smaller forLETd. In the obtainedLETddistribution, the deviation between the calculated and measured values agreed within 3% in the peak region, while the deviation was larger in other regions.Significance.The LET distribution can be measured with a single irradiation using two scintillator sheets. This method may be effective for verifying LET in daily clinical practice and for quality control.

Background.Adaptive radiotherapy (ART) requires precise tissue characterization to optimize treatment plans and enhance the efficacy of radiation delivery while minimizing exposure to organs at risk. Traditional imaging techniques such as cone beam computed tomography (CBCT) used in ART settings often lack the resolution and detail necessary for accurate dosimetry, especially in proton therapy.Purpose.This study aims to enhance ART by introducing an innovative approach that synthesizes dual-energy computed tomography (DECT) images from CBCT scans using a novel 3D conditional denoising diffusion probabilistic model (DDPM) multi-decoder. This method seeks to improve dose calculations in ART planning, enhancing tissue characterization.Methods.We utilized a paired CBCT-DECT dataset from 54 head and neck cancer patients to train and validate our DDPM model. The model employs a multi-decoder Swin-UNET architecture that synthesizes high-resolution DECT images by progressively reducing noise and artifacts in CBCT scans through a controlled diffusion process.Results.The proposed method demonstrated superior performance in synthesizing DECT images (High DECT MAE 39.582 ± 0.855 and Low DECT MAE 48.540± 1.833) with significantly enhanced signal-to-noise ratio and reduced artifacts compared to traditional GAN-based methods. It showed marked improvements in tissue characterization and anatomical structure similarity, critical for precise proton and radiation therapy planning.Conclusions.This research has opened a new avenue in CBCT-CT synthesis for ART/APT by generating DECT images using an enhanced DDPM approach. The demonstrated similarity between the synthesized DECT images and ground truth images suggests that these synthetic volumes can be used for accurate dose calculations, leading to better adaptation in treatment planning.

Objective. The acceleration of magnetic resonance imaging (MRI) acquisition is crucial for both clinical and research applications, particularly in dynamic MRI. Existing compressed sensing (CS) methods, despite being effective for fast imaging, face limitations such as the need for incoherent sampling and residual noise, which restrict their practical use for rapid MRI.Approach. To overcome these challenges, we propose a novel image reconstruction framework that integrates the MRI physical model with a flexible, self-adjusting, decoupling data-driven model. We validated this method through experiments using both simulated andin vivodynamic contrast-enhanced MRI datasets.Main results. The experimental results demonstrate that the proposed framework achieves high spatial and temporal resolution reconstructions. Additionally, when compared to state-of-the-art image reconstruction approaches, our method significantly enhances acceleration capabilities, enabling sparse and rapid imaging with high resolution.Significance. Our proposed framework offers a promising solution for real-time imaging and image-guided radiation therapy applications by providing superior performance in achieving high spatial and temporal resolution reconstructions, thus addressing the limitations of existing CS schemes.

Objective. MRI is the standard imaging modality for high-dose-rate brachytherapy of cervical cancer. Precise contouring of organs at risk (OARs) and high-risk clinical target volume (HR-CTV) from MRI is a crucial step for radiotherapy planning and treatment. However, conventional manual contouring has limitations in terms of accuracy as well as procedural time. To overcome these, we propose a deep learning approach to automatically segment OARs (bladder, rectum, and sigmoid colon) and HR-CTV from female pelvic MRI.Approach. In the proposed pipeline, a coarse multi-organ segmentation model first segments all structures, from which a region of interest is computed for each structure. Then, each organ is segmented using an organ-specific fine segmentation model separately trained for each organ. To account for variable sizes of HR-CTV, a size-adaptive multi-model approach was employed. For coarse and fine segmentations, we designed a dual convolution-transformer UNet (DCT-UNet) which uses dual-path encoder consisting of convolution and transformer blocks. To evaluate our model, OAR segmentations were compared to the clinical contours drawn by the attending radiation oncologist. For HR-CTV, four sets of contours (clinical + three additional sets) were obtained to produce a consensus ground truth as well as for inter/intra-observer variability analysis.Main results. DCT-UNet achieved dice similarity coefficient (mean ± SD) of 0.932 ± 0.032 (bladder), 0.786 ± 0.090 (rectum), 0.663 ± 0.180 (sigmoid colon), and 0.741 ± 0.076 (HR-CTV), outperforming other state-of-the-art models. Notably, the size-adaptive multi-model significantly improved HR-CTV segmentation compared to a single-model. Furthermore, significant inter/intra-observer variability was observed, and our model showed comparable performance to all observers. Computation time for the entire pipeline per subject was 12.59 ± 0.79 s, which is significantly shorter than the typical manual contouring time of >15 min.Significance. These experimental results demonstrate that our model has great utility in cervical cancer brachytherapy by enabling fast and accurate automatic segmentation, and has potential in improving consistency in contouring. DCT-UNet source code is available athttps://github.com/JHU-MICA/DCT-UNet.

Introduction: Ion therapies have an increased relative biological effectiveness (RBE) compared to X-rays, but this remains poorly quantified across different radiation qualities. Mechanistic models that simulate DNA damage and repair after irradiation could be used to help better quantify RBE. However, there is large variation in model design with the simulation detail and number of parameters required to accurately predict key biological endpoints remaining unclear. This work investigated damage models with varying detail to determine how different model features impact the predicted DNA damage. Methods: Damage models of reducing detail were designed in TOPAS-nBio and Medras investigating the inclusion of chemistry, realistic nuclear geometries, single strand break damage, and track structure. The nucleus models were irradiated with 1 Gy of protons across a range of linear energy transfers (LETs). Damage parameters in the models with reduced levels of simulation detail were fit to proton double strand break (DSB) yield predicted by the most detailed model. Irradiation of the optimised models with a range of radiation qualities was then simulated, before undergoing repair in the Medras biological response model. Results: Simplified damage models optimised to proton exposures predicted similar trends in DNA damage across radiation qualities. On average across radiation qualities, the simplified models experienced an 8% variation in double strand break (DSB) yield but a larger 28% variation in chromosome aberrations. Aberration differences became more prominent at higher LETs, with model features having an increasing impact on the distribution and therefore misrepair of DSBs. However, overall trends remained similar with better agreement likely achievable through repair model optimisation. Conclusion: Several model simplifications could be made without compromising key damage yield predictions, although changes in damage complexity and distribution were observed. This suggests simpler, more efficient models may be sufficient for initial radiation damage comparisons, if validated against experimental data. .

Objective. Optical fibre dosimeters (OFDs) offer great promise for real-timein vivodose measurement in radiation-based treatment modalities such as radiotherapy and brachytherapy. This is attributed to their many useful qualities such as high spatial resolution and sensitivity. However, there are several requirements that an optical fibre dosimeter must meet to be acceptable for dose measurement in a specified treatment modality. In this work, the dosimetric performance of a novel optical fibre dosimeter for use in external beam radiotherapy is presented.Approach. The dosimeter was characterised for photon beam energies between 6-15 MV using a Varian TrueBeam Linac at dose rates between 100-2400 MU/min and assessed based on its repeatability, dose dependence, dose rate dependence, energy dependence and dose-per-pulse dependence.Main Results. The results demonstrated excellent short-term repeatability of 0.3%, good linearity in response (R2>0.9997), and minor dose rate dependence between 0.53%-2.49% for all beam qualities investigated. As the scintillator of the OFD is non-water equivalent, Monte-Carlo-TOPAS simulations were used to calculate the absorbed dose energy dependence. A dose-per-pulse dependence was also investigated and compared with dosimetry measurements made with an ionisation chamber and simulated from the treatment planning system. An over-response of 20%was found at the lowest investigated dose-per-pulse, and an under-response of 34%was found at the highest investigated dose-per-pulse. This is believed to be due to an intrinsic energy dependence making this type of OFD unsuitable for external beam radiotherapy dosimetry.Significance. The OFD evaluated in this work was primarily designed for high-dose-rate brachytherapy whereas this study includes the first measurements made in external beam radiotherapy and highlights the challenges of transferability of the dosimeter to a different radiation source.

Cancer has a high incidence and lethality rate, which is a significant threat to human health. With the development of high-throughput technologies, different types of cancer genomics data have been accumulated, including genomics, epigenomics, transcriptomics, proteomics, and metabolomics. A comprehensive analysis of various omics data is needed to understand the underlying mechanisms of tumor development. However, integrating such a massive amount of data is one of the main challenges today. Artificial intelligence (AI) techniques such as machine learning are now becoming practical tools for analyzing and understanding multi-omics data on diseases. Enabling great optimization of existing research paradigms for cancer screening, diagnosis, and treatment. In addition, intelligent healthcare has received widespread attention with the development of healthcare informatization. As an essential part of innovative healthcare, practical, intelligent prognosis analysis and personalized treatment for cancer patients are also necessary. This paper introduces the advanced multi-omics data analysis technology in recent years, presents the cases and advantages of the combination of both omics data and AI applied to cancer diseases, and finally briefly describes the challenges faced by multi-omics analysis and AI at the current stage, aiming to provide new perspectives for oncology research and the possibility of personalized cancer treatment.