In computer-aided diagnosis systems, precise feature extraction from CT scans of colorectal cancer using deep learning is essential for effective prognosis. However, existing convolutional neural networks struggle to capture long-range dependencies and contextual information, resulting in incomplete CT feature extraction. To address this, the PEDRA-EFB0 architecture integrates patch embeddings and a dual residual attention mechanism for enhanced feature extraction and survival prediction in colorectal cancer CT scans. A patch embedding method processes CT scans into patches, creating positional features for global representation and guiding spatial attention computation. Additionally, a dual residual attention mechanism during the upsampling stage selectively combines local and global features, enhancing CT data utilization. Furthermore, this paper proposes a feature selection algorithm that combines autoencoders and entropy technology, encoding and compressing high-dimensional data to reduce redundant information and using entropy to assess the importance of features, thereby achieving precise feature selection. Experimental results indicate the PEDRA-EFB0 model outperforms traditional methods on colorectal cancer CT metrics, notably in C-index, BS, MCC, and AUC, enhancing survival prediction accuracy. Our code is freely available at https://github.com/smile0208z/PEDRA .

Chronic obstructive pulmonary disease (COPD) is a leading cause of death worldwide and greatly reduces the quality of life. Utilizing remote monitoring has been shown to improve quality of life and reduce exacerbations, but remains an ongoing area of research. We introduce a novel method for estimating changes in ease of breathing for COPD patients, using obstructed breathing data collected via wearables. Physiological signals were recorded, including respiratory airflow, acceleration, audio, and bio-impedance. By comparing patient-specific measurements, this approach enables non-intrusive remote monitoring. We analyze the influence of signal selection, window parameters, feature engineering, and classification models on predictive performance, finding that acceleration signals are most effective, complemented by audio signals. The best model achieves an F1-score of 0.83. To facilitate clinical adoption, we incorporate interpretability by designing novel saliency map methods, highlighting important aspects of the signals. We adapt local explainability techniques to time series and introduce a novel imputation method for periodic signals, improving faithfulness to the data and interpretability.

This study aims to understand the impact of backpack carriage, a regular activity for many, on back muscles and joint mobility during walking so that clinicians can develop strategies or products to ensure individuals’ safety and well-being. Surface electromyography (EMG) and XSENS Awinda motion capture systems were used to analyze the effects of carrying a backpack (12% of body weight) on erector spinae and multifidus muscles, as well as spinal, hip, knee, and ankle joints. Subjects walked at 4 km/h on flat and inclined surfaces. Paired t-tests compared backpack loads to baseline measurements. Carrying a backpack reduced activation levels in erector spinae and multifidus muscles and restricted spinal joint range of motion (axial and lateral bending, p<0.05). Hip joint rotation increased (p<0.05). Moderate to strong correlations were observed between muscle activity and spinal joint ROM, notably with left erector spinae and L5-S1 lateral bending (r=0.723,p<0.001). Backpack carriage decreases back muscle activation and alters the joint range of motion. Asymmetric correlations show that the subjects adapt muscle activity and gait patterns asymmetrically to manage external loads.Graphical abstract

Blood pressure (BP) is one of the vital physiological parameters, and its measurement is done routinely for almost all patients who visit hospitals. Cuffless BP measurement has been of great research interest over the last few years. In this paper, we aim to establish a method for cuffless measurement of BP using ultrasound. In this method, the arterial wall is pushed with an acoustic radiation force impulse (ARFI). After the completion of the ARFI pulse, the artery undergoes impulsive unloading which stimulates a hoop mode vibration. We designed two machine learning (ML) models which make it possible to estimate the internal pressure of the artery using ultrasonically measurable parameters. To generate the training data for the ML models, we did extensive finite element method (FEM) eigen frequency simulations for different tubes under pressure by sweeping through a range of values for inner lumen diameter (ILD), tube density (TD), elastic modulus, internal pressure (IP), tube length, and Poisson's ratio. Through image processing applied on images of different eigen modes supported for each simulated case, we identified its hoop mode frequency (HMF). Two different ML models were designed based on the simulated data. One is a four-parameter model (FPM) that takes tube thickness (TT), TD, ILD, and HMF as the inputs and gives out IP as output. Second is a three-parameter model (TPM) that takes TT, ILD, and HMF as inputs and IP as output. The accuracy of these models was assessed using simulated data, and their performance was confirmed through experimental verification on two arterial phantoms across a range of pressure values. The first prediction model (FPM) exhibited a mean absolute percentage error (MAPE) of 5.63% for the simulated data and 3.68% for the experimental data. The second prediction model (TPM) demonstrated a MAPE of 6.5% for simulated data and 8.73% for experimental data. We were able to create machine learning models that can measure pressure within an elastic tube through ultrasonically measurable parameters and verified their performance to be adequate for BP measurement applications. This work establishes a pathway for cuffless, continuous, real-time, and non-invasive measurement of BP using ultrasound.

IntroductionInfrared thermography (IRT) has gained significant interest in medical applications for its potential in diagnosing various conditions. Smartphone-based IRT modules offer portability and affordability, leading to increased utilization in medical settings. However, differences in performance among these modules raise questions about their reliability for medical use.Materials and methodsThis study compared three smartphone-based IRT modules (SmartIRT—Hikmicro, FLIR One Pro, and Seek Thermal CompactPRO—which, according to their datasheets, exhibit comparable quality and parameters. Temperature stability, surface temperature of the body, and spatial uniformity of provided images were assessed using calibrated black body measurements and surface temperature monitoring.ResultsThe Hikmicro module exhibited the most stable temperature readings, while FLIR One Pro showed the highest temperature increase over time. Seek Thermal CompactPRO demonstrated relatively better spatial uniformity. However, discrepancies in image resolution were noted, with FLIR One and Seek modules modifying image sizes through post-processing algorithms.ConclusionWhile SmartIRT modules offer affordability and portability, their performance varies significantly. Temporal stability emerges as a critical factor, with the Hikmicro module demonstrating leadership in this aspect. Careful consideration and validation are necessary when selecting and utilizing SmartIRT modules for medical applications.Graphical abstract

Type 2 diabetes (T2D) is becoming one of the leading health problems in Western societies, diminishing quality of life and consuming a significant share of healthcare resources. This study presents machine learning models for T2D diagnosis and prognosis, developed using heterogeneous data from a Spanish population dataset (Di@bet.es study). The models were trained exclusively on individuals classified as controls and undiagnosed diabetics, ensuring that the results are not influenced by treatment effects or behavioral changes due to disease awareness. Two data domains are considered: environmental (patient lifestyle questionnaires and measurements) and clinical (biochemical and anthropometric measurements). The preprocessing pipeline consists of four key steps: geospatial data extraction, feature engineering, missing data imputation, and quasi-constancy filtering. Two working scenarios (Environmental and Healthcare) are defined based on the features used, and applied to two targets (diagnosis and prognosis), resulting in four distinct models. The feature subsets that best predict the target have been identified based on permutation importance and sequential backward selection, reducing the number of features and, consequently, the cost of predictions. In the Environmental scenario, models achieved an AUROC of 0.86 for diagnosis and 0.82 for prognosis. The Healthcare scenario performed better, with an AUROC of 0.96 for diagnosis and 0.88 for prognosis. A partial dependence analysis of the most relevant features is also presented. An online demo page showcasing the Environmental and Healthcare T2D prognosis models is available upon request.Graphic abstractSupplementary InformationThe online version contains supplementary material available at 10.1007/s11517-025-03355-5.

Ventricular cerebrospinal fluid (CSF) flow has bi-directional flow profiles synchronized with cardiac pulsation. The increase in CSF stroke volume through the aqueduct (flow pathway between the third and fourth ventricle) is believed to be a biomarker of idiopathic normal pressure hydrocephalus (iNPH). However, several studies have reported that CSF stroke volume varies considerably even in healthy populations. To explore these variations from a fluid mechanics perspective, this study analyzed CSF flow characteristics in healthy individuals using magnetic resonance imaging (MRI)-based computational simulations. MRI data from 47 healthy subjects were acquired, and the maximum Reynolds number of the CSF flow through the aqueduct and the degree of CSF mixing were evaluated. Results showed that the Reynolds number of the CSF flow through the aqueduct was 28.6 ± 13.3, and the limited variation suggested fluid mechanical similarities of CSF flow characteristics in the healthy population. A positive correlation between CSF flow mixing and the Reynolds number was observed in both healthy populations and iNPH patients. These findings demonstrate that CSF flow through the aqueduct, especially in healthy populations, can be well characterized by examining fluid mechanical similarities.Graphical

The demand for total hip replacement surgery is increasing year by year. However, the issue of hip prosthesis failure, particularly the modular acetabular cup, still exists. The performance and functional requirements of modular acetabular cups have not yet met clinical expectations. This study focused on poly-ether-ether-ketone (PEEK) shells, using finite element methods to investigate their mechanical stability under gait loads and motion, including parameters such as deformation, micromotion, and bone strain. The results showed that a compromise was required among the mechanical performance, stability, and bone integration capabilities of the PEEK shell. As the shell rigidity increased, deformation decreased. However, increased rigidity also increased micromotion at the bone-prosthesis interface, reducing the area that promoted bone ingrowth. Additionally, potential bone absorption areas were also increased, reducing bone preservation and reconstruction capabilities. Compromises need to be made among mechanical performance, stability, and bone integration to achieve optimal mechanical stability. In this study, a 6mm wall thickness PEEK shell was found to provide good overall mechanical stability.

Echocardiography is a primary tool for cardiac diagnosis. Accurate delineation of the left ventricle is a prerequisite for echocardiography-based clinical decision-making. In this work, we propose an echocardiographic left ventricular segmentation method based on the diffusion probability model, which is named EchoSegDiff. The EchoSegDiff takes an encoder-decoder structure in the reverse diffusion process. A diffusion encoder residual block (DEResblock) based on the atrous pyramid squeeze attention (APSA) block is coined as the main module of the encoder, so that the EchoSegDiff can catch multiscale features effectively. A novel feature fusion module (FFM) is further proposed, which can adaptively fuse the features from encoder and decoder to reduce semantic gap between encoder and decoder. The proposed EchoSegDiff is validated on two publicly available echocardiography datasets. In terms of left ventricular segmentation performance, it outperforms other state-of-the-art networks. The segmentation accuracy on the two datasets reached 93.69% and 89.95%, respectively. This demonstrates the excellent potential of EchoSegDiff in the task of left ventricular segmentation in echocardiography.

Predicting tumor biomarkers with high precision is essential for improving the diagnostic accuracy and developing more effective treatment strategies. This paper proposes a machine learning model that utilizes CT images and biopsy whole slide images (WSI) to classify mesothelin expression levels in pancreatic cancer. By combining multimodal learning and stochastic configuration networks, a radiopathomics mesothelin-prediction system named RPMSNet is developed. The system extracts radiomic and pathomic features from CT images and WSI, respectively, and sends them into stochastic configuration networks for the final prediction. Compared to traditional radiomics or pathomics, this system has the capability to capture more comprehensive image features, providing a multidimensional insight into tissue characteristics. The experiments and analyses demonstrate the accuracy and effectiveness of the system, with an area under the curve of 81.03%, an accuracy of 73.67%, a sensitivity of 71.54%, a precision of 76.78%, and a F1-score of 72.61%, surpassing both single-modality and dual-modality models. RPMSNet highlights its potential for early diagnosis and personalized treatment in precision medicine.