Biomedical Physics & Engineering Express最新文献

In-ear electroencephalography (EEG) systems offer several practical advantages over scalp-based EEG systems for non-invasive brain-computer interface (BCI) applications. However, the difficulty in fabricating in-ear EEG systems can limit their accessibility for BCI use cases. In this study, we developed a portable, low-cost wireless in-ear EEG device using commercially available components. In-ear EEG signals (referenced to left mastoid) from 5 adolescent participants were compared to scalp-EEG collected simultaneously during an alpha modulation task, various artifact induction tasks, and an auditory word-streaming BCI paradigm. Spectral analysis confirmed that the proposed in-ear EEG system could capture significantly increased alpha activity during eyes-closed relaxation in 3 of 5 participants, with a signal-to-noise ratio of 2.34 across all participants. In-ear EEG signals were most susceptible to horizontal head movement, coughing and vocalization artifacts but were relatively insensitive to ocular artifacts such as blinking. For the auditory streaming paradigm, the classifier decoded the presented stimuli from in-ear EEG signals only in 1 of 5 participants. Classification of the attended stream did not exceed chance levels. Contrast plots showing the difference between attended and unattended streams revealed reduced amplitudes of in-ear EEG responses relative to scalp-EEG responses. Hardware modifications are needed to amplify in-ear signals and measure electrode-skin impedances to improve the viability of in-ear EEG for BCI applications.

The rapid development of deep learning-based computational pathology and genomics has demonstrated the significant promise of effectively integrating whole slide images (WSIs) and genomic data for cancer survival prediction. However, the substantial heterogeneity between pathological and genomic features makes exploring complex cross-modal relationships and constructing comprehensive patient representations challenging. To address this, we propose the Information Compression-based Multimodal Confidence-guided Fusion Network (iMCN). The framework is built around two key modules. First, the Adaptive Pathology Information Compression (APIC) module employs learnable information centers to dynamically cluster image regions, removing redundant information while maintaining discriminative survival-related patterns. Second, the Confidence-guided Multimodal Fusion (CMF) module utilizes a learned sub-network to estimate the confidence of each modality's representation, allowing for dynamic weighted fusion that prioritizes the most reliable features in each case. Evaluated on the TCGA-LUAD and TCGA-BRCA cohorts, iMCN achieved average concordance index (C-index) values of 0.691 and 0.740, respectively, outperforming existing state-of-the-art methods by an absolute improvement of 1.65%. Qualitatively, the model generates interpretable heatmaps that localize high-association regions between specific morphological structures (e.g., tumor cell nests) and functional genomic pathways (e.g., oncogenesis), offering biological insights into genomic-pathologic linkages. In conclusion, iMCN significantly advances multimodal survival analysis by introducing a principled framework for information compression and confidence-based fusion. Besides, correlation analysis reveal that tissue heterogeneity influences optimal retention rates differently across cancer types, with higher-heterogeneity tumors (e.g., LUAD) benefiting more from aggressive information compression. Beyond its predictive performance, the model's ability to elucidate the interplay between tissue morphology and molecular biology enhances its value as a tool for translational cancer research.

Efficient and accurate image segmentation models play a vital role in medical image segmentation, however, high computational cost of traditional models limits clinical deployment. Based on pyramid visual transformers and convolutional neural networks, this paper proposes a lightweight Context Contrast Enhancement Network (CCE-Net) that ensures efficient inference and achieves accurate segmentation through the contextual feature synergy mechanism and feature contrast enhancement strategy. The Local Context Fusion Enhancement module is designed to obtain more specific local detail information through cross-layer context fusion and bridge the semantic gap between the encoder and decoder. The Deep Feature Multi-scale Extraction module is proposed to fully extract the comprehensive information about the deepest features in the bottleneck layer of the model and provide more accurate global contextual features for the decoder. The Detail Contrast Enhancement Decoder module is designed to effectively solve the inherent problems of missing image details and blurred edges through adaptive dual-branch feature fusion and frequency-domain contrast enhancement operations. Experiments show that CCE-Net only requires 5.40M parameters and 0.80G FLOPs, and the average Dice coefficients on the Synapse and ACDC datasets are 82.25% and 91.88%, respectively, which are 37%-62% less than the parameters of mainstream models, promoting the transformation of lightweight medical AI models from laboratory research to clinical practice.

Objective.In the revised version of the TRS-398 Code of Practice (CoP), Monte Carlo (MC) results were added to existing experimental data to derive the recommended beam quality correction factors (k_Q) for ionisation chambers in proton beams. While part of these results were obtained from versions v10.3 and v10.4 of the Geant4 simulation tool, this paper demonstrates that the use of a more recent version, such as v11.2, can affect the value of thek_Qfactors.Approach.The chamber-specific proton contributions (f_Q) of thek_Qfactors were derived for four ionisation chambers using two different versions of the code, namely Geant4-v.10.3 and Geant4-v11.2. A comparison of the total absorbed dose values is performed, as well as the comparison of the dose contribution for primary and secondary particles.Main results.Larger absorbed dose values per incident particle were derived with Geant4-v11.2 compared to Geant4-v10.3 especially for dose-to-air at high proton beam energies between 150 MeV and 250 MeV, leading to deviations in thek_Qvalues up to 1%. These deviations are mainly due to a change in the physics of secondary helium ions for which the significant deviations between the Geant4 versions is the most stringent within the entrance window or the shell of the ionisation chambers.Significance.Although significant deviations in the MC calculatedf_Qvalues were observed between the two Geant4 versions, the dominant uncertainty of theW_airvalues currently allows to achieve the agreement at thek_Qlevel. As these values also agree with the current data presented in the TRS-398 CoP, it is not possible at the moment to discriminate between Geant4-v10.3 and Geant4-v11.2, which are therefore both suitable fork_Qcalculation.

Knee Osteoarthritis (KOA) is a prevalent degenerative joint disease affecting millions worldwide. Accurate classification of KOA severity is crucial for effective diagnosis and treatment planning. This study introduces a novel multi-classification algorithm for x-ray KOA based on MambaOut and Latent Diffusion Model (LDM). MambaOut, an emerging network architecture, achieves superior classification performance compared to fine-tuning the mainstream Convolutional Neural Networks (CNNs) for KOA classification. To address sample imbalance across KL grades, we propose an AI-generated model using LDM to synthesize new data. This approach enhances minority-class samples by optimizing the autoencoder's loss function and incorporating pathological labels into the LDM framework. Our approach achieves an average accuracy of 86.3%, an average precision of 85.3%, an F1 score of 0.855, and a mean absolute error reduced to 14.7% in the four-classification task, outperforming recent advanced methods. This study not only advances KOA classification techniques but also highlights the potential of integrating advanced neural architectures with generative models for medical image analysis.

Objective: In radioembolization, SPECT/CT planning scans are often acquired during free breathing, which can introduce motion-related blurring and misregistration between SPECT and CT, leading to dosimetric inaccuracies. This study quantifies the impact of respiratory motion on absorbed dose metrics-tumor-to-normal tissue (T/N) ratio, dose volume histograms, and mean dose-using several voxel-based dosimetry methods. This study supports standardization efforts through experimental measurements using a motion-enabled phantom. Approach: Motion effects in pre-therapy imaging were evaluated using a Jaszczak phantom filled with technetium-99m, simulating activity in lesion and background volumes. SPECT/CT scans were acquired with varying cranial-caudal motion amplitudes from the central position ±0, ±5, ±6.5, ±10, ±12.5, and ±15 mm. The impact of motion-related misregistration during scanning on dosimetry was also examined. Five dosimetry methods, including Monte Carlo simulation with uniform reference activity (MC REF), Monte Carlo simulation based on SPECT images (MC SPECT), Simplicity™ (Boston Scientific), local deposition method, and voxel-S-value convolution. Absorbed dose metrics of mean dose, dose volume histogram dosimetric indices (D50, D70, D90), and T/N ratio were obtained to quantify motion effects and evaluate clinical suitability. Main results: Mean absorbed dose values for the lesion and background were consistent across methods within uncertainties, though discrepancies were noted in non-lesion low-density regions. Respiratory motion reduced lesion dose by 16-25% and increased background dose by 13-32%, although the latter represented only a 1-2 Gy change. These shifts led to a 28-43% decrease in the T/N ratio at ±12.5 mm motion amplitude. Misregistration due to motion also significantly impacted dosimetric accuracy. Significance: The study demonstrated agreement between five dosimetry methods and revealed that respiratory motion can lead to substantial underestimation of the lesion dose and T/N ratio. Since T/N ratio is critical for patient selection and activity prescription, accounting for respiratory motion is essential for accurate radioembolization dosimetry.

Electronic health records (EHRs) capture longitudinal multi-visit patient journeys but are difficult to analyze due to temporal irregularity, multimorbidity, and heterogeneous coding. This study introduces a temporal and comorbidity-aware trajectory representation that restructures admissions into ordered symbolic visit states while preserving diagnostic progression, secondary comorbidities, procedure categories, demographics, outcomes, and inter-visit intervals. These symbolic states are subsequently encoded as fixed-length numerical vectors suitable for computational analysis. Validation was conducted in two stages: Stage I assessed construction fidelity using coverage metrics, comorbidity preservation, diagnostic transition structures, and exact inter-visit gap encoding and Stage II assessed analytical utility through clustering experiments using different clustering approacheslike sequence similarity, Gaussian Mixture Models (GMM), and a temporal LSTM autoencoder (TS-LSTM). Proof of concept was done by encoding subset of patient cohorts from the MIMIC-IV database consisting of 2,280 patients with 8,849 admissions having complete primary diagnosis coverage and near-complete secondary coverage. Stage 1 assessment consisting of cohort-level coverage metrics confirmed that the transformation preserved essential clinical information and key properties of longitudinal EHRs. In Stage 2, clustering experiments validated the analytical utility of the representation across sequence-based, Gaussian mixture, and temporal LSTM autoencoder approaches. Ablation studies further demonstrated that both multimorbidity depth and inter-visit gap encoding are critical to maintaining cluster separability and temporal fidelity. The findings show that explicit encoding of comorbidity and timing improves interpretability and subgroup coherence. Although evaluated on a single dataset, the use of standardised ICD-10 EHR structure supports the assumption that the framework can generalise across healthcare settings; future work will incorporate multimodal data and external validation.

Research on deep learning for medical image segmentation has shifted from single-modality networks to multimodal data fusion. Updating the parameters of such deep learning models is crucial for accurate segmentation predictions. Although existing optimizers can perform global parameter updates, the fine-grained initialization of learning rates across different network hierarchies and its influence on segmentation performance has not been sufficiently explored. To address this, we conducted a series of experiments showing that the initialization of a differentiated learning rate across network layers directly affected the performance of medical image segmentation models. To determine the optimal initial learning rate for each network level, we summarized a general statistical relationship between early-stage training results and the model's final optimal performance. In this paper, we proposed a fine-grained learning rate configuration algorithm. To verify the effectiveness of the proposed algorithm, we evaluated 10 segmentation models on three benchmark datasets: the colon polyp segmentation dataset CVC-ClinicDB, the gastrointestinal polyp dataset Kvasir-SEG, and the breast tumor segmentation dataset BUSI. The models that achieved the most significant improvement in mIoU on these three datasets were H-vmunet, MSRUNet, and H-vmunet, with increases of 3.87%, 4.67%, and 6.22%, respectively. Additionally, we validated the generalization and transferability of the proposed algorithm using a thyroid nodule segmentation dataset and a skin lesion segmentation dataset. Finally, a series of analyses, including segmentation result analysis, feature map visualization, training process analysis, computational overhead analysis, and clinical relevance analysis, confirmed the effectiveness of the proposed method. The core code is publicly available athttps://github.com/Lambda-Wave/PaperCoreCode.

Purpose. Mammographic density is associated with the risk of developing breast cancer and can be predicted using deep learning methods. Model uncertainty estimates are not produced by standard regression approaches but would be valuable for clinical and research purposes. Our objective is to produce deep learning models with in-built uncertainty estimates without degrading predictive performance.Approach. We analysed data from over 150,000 mammogram images with associated continuous density scores from expert readers in the Predicting Risk Of Cancer At Screening (PROCAS) study. We re-designated the continuous density scores to 100 density classes then trained classification and ordinal deep learning models. Distributions and distribution-free methods were applied to extract predictions and uncertainties. A deep learning regression model was trained on the continuous density scores to act as a direct comparison.Results. The root mean squared error (RMSE) between expert assigned density labels and predictions of the standard regression model were 8.42 (8.34-8.51) while the RMSE for the classification and ordinal classification were 8.37 (8.28-8.46) and 8.44 (8.35-8.53) respectively. The average uncertainties produced by the models were higher when the density scores from pairs of expert readers density scores differ more, when different mammogram views of the same views are more variable, and when two separately trained models show higher variation.Conclusions. Using either a classification or ordinal approach we can produce model uncertainty estimates without loss of predictive performance.

Objective. Proton boron capture therapy (PBCT) is a novel approach that utilizes alpha particles generated through the proton induced capture reaction with¹¹B. Early studies reported substantial dose enhancements of 50%-96% near the Bragg peak, suggesting a promising therapeutic advantage. However, subsequent investigations have raised critical concerns regarding the practical feasibility of PBCT, primarily due to the relatively low reaction cross section in the Bragg peak region and the need for clinically unrealistic boron concentrations. The aim of this study is to evaluate Relative Biological Effectiveness (RBE) enhancement in PBCT using microdosimetric analysis across a wide range of boron concentrations.Approach. In the present work, we have employed Monte Carlo model using Particle and Heavy Ion Transport code System (PHITS) package combined with the Microdosimetric Kinetic Model to quantify both physical and biological dose enhancements at varying concentrations of¹¹B.Main Results. Microdosimetric analysis revealed that the total dose is dominated by protons, although alpha particles dominate in regions of higher linear energy deposition. The resulting RBE enhancement factors were 1.0011, 1.0080, and 1.1275 (for Human Salivary Gland (HSG) cell type) for 100, 1000, and 10,000 ppm boron concentrations, respectively. While the enhancements at lower concentrations are negligible, a modest increase is observed at very high boron levels.Significance. Based on the resulting RBE enhancement factors, it can be concluded that although alpha particles generated via thep+¹¹B → 3αreaction contribute high-Linear Energy Transfer (LET) energy at the cellular level, the overall biological dose enhancement remains rather minimal. These results indicate that under clinically achievable boron concentrations, the therapeutic benefit of PBCT may be limited.