Hasil untuk "Computer applications to medicine. Medical informatics"

Poroshista Knauer, Gernot Steiner, Rony-Orijit Dey Hazra et al.

Abstract Background Range of motion (ROM) measurements are performed every day in clinical practice. The ROM is usually measured using conventional goniometers; however, studies have shown considerable variability in interrater and intrarater reliability. Despite the growing availability of smartphone-based goniometry, studies on the usability of these tools are highly limited. The aim of this study was to evaluate shoulder ROM applications (apps), with a focus on their usability and handling. Methods Eleven apps were identified and tested by two physicians using the same smartphone. A healthy volunteer subject performed defined movements with each arm measured three times per app by both users. The apps were evaluated in terms of usability on a grading scale, with 1 representing the best result and 6 representing the worst result, on the basis of two key factors: intuitiveness and quality of description. The learning time of the physicians was also measured and recorded in minutes. Results The usability of the apps showed considerable variability. The usability is statistically correlated with the learning time of the raters (p = 0.021; τb = 0.607 for quality of description and p = 0.039; τb = 0.551 for intuitiveness). The quality of description and intuitiveness are significantly correlated with each other (p < 0.001; τb = 0.941). The interrater reliability was high, with an ICC of 0.88 for the quality of the description and 0.95 for intuitiveness, suggesting consistent evaluations between the raters. Two apps were rated as the best apps, with 1.0 and 1.25 in usability. Conclusion Overall, there seems to be wide variation in the usability of smartphone-based apps. While smartphone-based goniometers show potential benefits, further research is required to establish their applicability before routine use.

Saul Martinez-Horta, Angela Quevedo-García, Arnau Puig-Davi et al.

Background: Huntington’s disease (HD) is primarily associated with executive dysfunction, but episodic memory impairment is also present. Traditionally, these memory deficits have been attributed to retrieval difficulties linked to fronto-striatal dysfunction, rather than to disruptions in encoding or consolidation processes. However, the specific nature and diversity of memory impairments in HD remain underexplored. Objective: To characterize the profile of episodic memory impairment in HD, identify distinct cognitive phenotypes, and examine their clinical, neuroanatomical, and biomarker correlates. Methods: We assessed episodic memory in HD patients and healthy controls using the Free and Cued Selective Reminding Test (FCSRT), complemented by Item-Specific Deficit Approach (ISDA) indices to quantify encoding, consolidation, and retrieval deficits. Structural MRI was used to identify gray matter volume correlates, and plasma neurofilament light chain (NfL) was measured as a marker of neuroaxonal injury. Results: Compared to controls, HD patients showed marked impairments in free recall with preserved cued recall, suggesting predominant retrieval deficits. However, nearly one-third of patients exhibited global impairments across all FCSRT components, mainly driven by consolidation deficits consistent with medial temporal lobe dysfunction. This subgroup also showed worse cognitive and functional performance and significant atrophy in the hippocampus, entorhinal cortex, and parahippocampal gyrus. Conclusion: Episodic memory dysfunction in HD is heterogeneous and includes both retrieval-related and consolidation-driven profiles. These profiles reflect distinct neurodegenerative patterns, emphasizing the importance of cognitive subtyping for improving clinical characterization and biomarker development in HD.

Sebastian Amaya, Sidhant Kalsotra, Sibelle Aurelie Yemele Kitio et al.

Abstract BackgroundThe perioperative environment is complex and may be challenging for patients and guardians to navigate. The emotional burden and stressors inherent to the perioperative process commonly result in preoperative anxiety. Many studies have demonstrated the usefulness of virtual reality (VR) in various patient populations. ObjectiveThe aim of this study is to evaluate the impact of a VR-based preoperative education tool on anxiety levels in pediatric patients undergoing ambulatory ear, nose, and throat surgery, as well as in their guardians. MethodsWe performed a single-center prospective randomized controlled trial including children 6‐12 years of age, presenting for ambulatory tonsillectomy and/or adenoidectomy, with or without bilateral ear tube insertion. The patients were randomized to receive VR instruction of the perioperative workflow or standard preoperative experience (non-VR). The primary outcome was patient and guardian preoperative anxiety, as measured by the 6-item State-Trait Anxiety Inventory. ResultsThe study cohort included 107 patient-guardian dyads—51 in the intervention (VR) group and 56 in the control (non-VR) group. Baseline characteristics between the study and control groups were comparable; however, patients in the control group were more likely to report feeling upset compared to the VR group. The VR intervention was associated with reduced preoperative anxiety in patients and guardians compared to the control group. Patients exposed to the VR intervention had higher odds of feeling calm (OR 4.95, 95% CI 2.32‐10.61; PPPPPP ConclusionsVR exposure may be effective in reducing pediatric and guardian anxiety. VR may be a suitable alternative to pharmacologic anxiolysis and future studies should compare the effect to premedication techniques.

Talshyn Sarsembayeva, Madina Mansurova, Assel Abdildayeva et al.

The accurate segmentation of lung regions in computed tomography (CT) scans is critical for the automated analysis of lung diseases such as chronic obstructive pulmonary disease (COPD) and COVID-19. This paper focuses on enhancing the accuracy of U-Net segmentation models through a robust preprocessing pipeline. The pipeline includes CT image normalization, binarization to extract lung regions, and morphological operations to remove artifacts. Additionally, the proposed method applies region-of-interest (ROI) filtering to isolate lung areas effectively. The dataset preprocessing significantly improves segmentation quality by providing clean and consistent input data for the U-Net model. Experimental results demonstrate that the Intersection over Union (IoU) and Dice coefficient exceeded 0.95 on training datasets. This work highlights the importance of preprocessing as a standalone step for optimizing deep learning-based medical image analysis.

Johnny Joyce, Jan Verschelde

Neural networks are important tools in machine learning. Representing piecewise linear activation functions with tropical arithmetic enables the application of tropical geometry. Algorithms are presented to compute regions where the neural networks are linear maps. Through computational experiments, we provide insights on the difficulty to train neural networks, in particular on the problems of overfitting and on the benefits of skip connections.

Ujjwal Mishra, Vinita Shukla, Praful Hambarde et al.

Adapting Vision Language Segmentation Models (VLSMs) to medical imaging domains requires significant computational overhead when using conventional fine-tuning approaches. Existing Parameter-Efficient Fine-Tuning (PEFT) methods apply uniform adapter dimensions across all transformer layers, leading to suboptimal parameter allocation and reduced adaptation efficiency. We introduce Telescopic Adapters, a novel PEFT framework that employs depth-aware scaling to progressively increase adapter capacity from shallow to deep transformer layers. Our method integrates lightweight bottleneck modules within CLIPSeg's vision and text encoders, with adapter dimensions dynamically scaled based on layer depth and semantic relevance. Using only 613k trainable parameters--244x fewer than end-to-end fine-tuning, Telescopic Adapters achieve superior performance across five diverse medical datasets spanning polyp segmentation, skin lesion detection, and breast ultrasound imaging. Comprehensive ablation studies demonstrate that deeper layers require substantially more adaptation capacity than shallow layers, validating our telescopic scaling hypothesis. Our approach establishes a new paradigm for efficient medical VLSM fine-tuning, enabling deployment in resource-constrained clinical environments while maintaining competitive segmentation accuracy. Our source code is publicly available at https://github.com/Ujjwal238/Telescopic_adapters

Arti Virkud, Jessie K. Edwards, Michele Jonsson Funk et al.

Identifying optimal medical treatments to improve survival has long been a critical goal of pharmacoepidemiology. Traditionally, we use an average treatment effect measure to compare outcomes between treatment plans. However, new methods leveraging advantages of machine learning combined with the foundational tenets of causal inference are offering an alternative to the average treatment effect. Here, we use three unique, precision medicine algorithms (random forests, residual weighted learning, efficient augmentation relaxed learning) to identify optimal treatment rules where patients receive the optimal treatment as indicated by their clinical history. First, we present a simple hypothetical example and a real-world application among heart failure patients using Medicare claims data. We next demonstrate how the optimal treatment rule improves the absolute risk in a hypothetical, three-modifier setting. Finally, we identify an optimal treatment rule that optimizes the time to outcome in a real-world heart failure setting. In both examples, we compare the average time to death under the optimized, tailored treatment rule with the average time to death under a universal treatment rule to show the benefit of precision medicine methods. The improvement under the optimal treatment rule in the real-world setting is greatest (additional ~9 days under the tailored rule) for survival time free of heart failure readmission.

Laura Hanke, Richard Schulte, Christian Boedecker et al.

BackgroundWorking in an operating room (OR) is physically and mentally challenging: the operation itself demands the surgeon's full attention, while time and cost efficiency constraints, daily planning, and emergency care interfere with the procedure. Thus, multitasking becomes an integral surgical competence. This study aims to examine the effect of disruptions during surgery in a highly immersive virtual reality (IVR) operation environment combined with a virtual reality (VR) laparoscopy simulator. ObjectiveThis study aims to identify distractions in the OR and their importance in the clinical setting. MethodsAn IVR environment was created using a high-resolution, stereoscopic 360° video of the OR. Different distractions were identified, classified as auditory, visual, or audio-visual, and recorded accordingly. The surrounding was combined with a VR laparoscopic simulator. Participants—medical students and surgical residents—received proficiency-based training in basic laparoscopic skills and were blinded to the aim of the experiment. Following a cross-over design, each participant received a unique order of virtual distraction factors while performing tasks on the laparoscopic simulator. During the experiment, subjective passing of time, stress, heart rate, and visually induced motion sickness are recorded. After the experiment, validated questionnaires for usability, immersion, and stress were completed, as well as subjective evaluation of the distractions. The questionnaires used included the system usability scale, Self-Assessment Manikin score, National Aeronautics and Space Administration Task Load Index, and the immersion rating scale as described by Nichols. Performance in the laparoscopic tasks in relation to distractions will be evaluated by the Wilcoxon test and ANOVA for continuous variables. Subgroup analyses in regard to age, gender, and expertise (medical students vs surgical residents) are planned. ResultsThe described trial started in August 2022 and is ongoing. By July 2024, a total of 30 medical students and 9 surgeons have completed the study. ConclusionsWe present a study protocol aiming to identify the impact of different disruptions in OR during laparoscopic training in IVR. Hence, it may lead to an improved awareness of distractions and facilitate accommodations toward an improved work environment. Prior research leads to the hypothesis that the performance of a more experienced surgeon is less impacted by distractions than the performance of inexperienced surgeons and medical students. Furthermore, we investigate which type of distraction has the largest impact on performance. With this knowledge, specific multitasking training can be devised, which may be particularly useful in medical education, for which VR might play a leading role. Additionally, workplace surroundings in the OR can be optimized with this knowledge. Trial RegistrationGerman Registry for Clinical Trials DRKS00030033; https://drks.de/search/en/trial/DRKS00030033 International Registered Report Identifier (IRRID)DERR1-10.2196/59014

Ur Metser, Andres Kohan, Catherine O’Brien et al.

Tumor hypoxia is a negative prognostic factor in many tumors and is predictive of metastatic spread and poor responsiveness to both chemotherapy and radiotherapy. <b>Purpose:</b> To assess the feasibility of using ¹⁸F-Fluoroazomycin arabinoside (FAZA) PET/MR to image tumor hypoxia in patients with locally advanced rectal cancer (LARC) prior to and following neoadjuvant chemoradiotherapy (nCRT). The secondary objective was to compare different reference tissues and thresholds for tumor hypoxia quantification. <b>Patients and Methods:</b> Eight patients with histologically proven LARC were included. All patients underwent ¹⁸F-FAZA PET/MR prior to initiation of nCRT, four of whom also had a second scan following completion of nCRT and prior to surgery. Tumors were segmented using T₂-weighted MR. Each voxel within the segmented tumor was defined as hypoxic or oxic using thresholds derived from various references: ×1.0 or ×1.2 SUVmean of blood pool [BP] or left ventricle [LV] and SUVmean +3SD for gluteus maximus. Correlation coefficient (CoC) between HF and tumor SUVmax/reference SUVmean TRR for the various thresholds was calculated. Hypoxic fraction (HF), defined as the % hypoxic voxels within the tumor volume was calculated for each reference/threshold. <b>Results:</b> For all cases, baseline and follow-up, the CoCs for gluteus maximus and for BP and LV (×1.0) were 0.241, 0.344, and 0.499, respectively, and HFs were (median; range) 16.6% (2.4–33.8), 36.8% (0.3–72.9), and 30.7% (0.8–55.5), respectively. For a threshold of ×1.2, the CoCs for BP and LV as references were 0.611 and 0.838, respectively, and HFs were (median; range) 10.4% (0–47.6), and 4.3% (0–20.1%), respectively. The change in HF following nCRT ranged from (−18.9%) to (+54%). <b>Conclusions:</b> Imaging of hypoxia in LARC with ¹⁸F-FAZA PET/MR is feasible. Blood pool as measured in the LV appears to be the most reliable reference for calculating the HF. There is a wide range of HF and variable change in HF before and after nCRT.

Jiajie Zhou, Xiaoming Yuan

Abstract Introduction Incarceration occurred in approximately 5% to 15% of inguinal hernia patients, with around 15% of incarcerated cases progressing to intestinal necrosis, necessitating bowel resection surgery. Patients with intestinal necrosis had significantly higher mortality and complication rates compared to those without necrosis.The primary objective of this study was to design and validate a diagnostic model capable of predicting intestinal necrosis in patients with incarcerated groin hernias. Methods We screened the clinical records of patients who underwent emergency surgery for incarcerated inguinal hernia between January 1, 2015, and December 31, 2022. To ensure balanced representation, the enrolled patients were randomly divided into a training set (n = 180) and a validation set (n = 76) using a 2:1 ratio. Logistic regression analysis was conducted using the rms package in R software, incorporating selected features from the LASSO regression model, to construct a predictive model. Results Based on the results of the LASSO regression analysis, a multivariate logistic regression model was developed to establish the predictive model. The predictors included in the model were Abdominal effusion, Hernia Sac Effusion, and Procalcitonin. The area under the receiver operating characteristic (ROC) curve for the nomogram graph in the training set was 0.977 (95% CI = 0.957–0.992). In the validation set, the AUC for the nomogram graph was 0.970. Calibration curve and decision curve analysis (DCA) verified the accuracy and practicability of the nomogram graph in our study. Conclusion Bowel necrosis in patients with incarcerated inguinal hernia was influenced by multiple factors. The nomogram predictive model constructed in this study could be utilized to predict and differentiate whether incarcerated inguinal hernia patients were at risk of developing bowel necrosis.

Duong Dang, Tero Vartiainen, Mike Mekkanen

Literature in cyber security including cyber security in energy informatics are tecnocentric focuses that may miss the chances of understanding a bigger picture of cyber security measures. This research thus aims to conduct a literature review focusing on non-technical issues in cyber security in the energy informatics field. The findings show that there are seven non-technical issues have been discussed in literature, including education, awareness, policy, standards, human, and risks, challenges, and solutions. These findings can be valuable for not only researchers, but also managers, policy makers, and educators.

Andrea Posada, Daniel Rueckert, Felix Meissen et al.

Since the Transformer architecture emerged, language model development has grown, driven by their promising potential. Releasing these models into production requires properly understanding their behavior, particularly in sensitive domains like medicine. Despite this need, the medical literature still lacks practical assessment of pre-trained language models, which are especially valuable in settings where only consumer-grade computational resources are available. To address this gap, we have conducted a comprehensive survey of language models in the medical field and evaluated a subset of these for medical text classification and conditional text generation. The subset includes 53 models with 110 million to 13 billion parameters, spanning the Transformer-based model families and knowledge domains. Different approaches are employed for text classification, including zero-shot learning, enabling tuning without the need to train the model. These approaches are helpful in our target settings, where many users of language models find themselves. The results reveal remarkable performance across the tasks and datasets evaluated, underscoring the potential of certain models to contain medical knowledge, even without domain specialization. This study thus advocates for further exploration of model applications in medical contexts, particularly in computational resource-constrained settings, to benefit a wide range of users. The code is available on https://github.com/anpoc/Language-models-in-medicine.

Niklas Gunnarsson, Jens Sjölund, Peter Kimstrand et al.

Image monitoring and guidance during medical examinations can aid both diagnosis and treatment. However, the sampling frequency is often too low, which creates a need to estimate the missing images. We present a probabilistic motion model for sequential medical images, with the ability to both estimate motion between acquired images and forecast the motion ahead of time. The core is a low-dimensional temporal process based on a linear Gaussian state-space model with analytically tractable solutions for forecasting, simulation, and imputation of missing samples. The results, from two experiments on publicly available cardiac datasets, show reliable motion estimates and an improved forecasting performance using patient-specific adaptation by online learning.

Max Menssen, Martina Dammann, Firas Fneish et al.

In pre-clinical and medical quality control, it is of interest to assess the stability of the process under monitoring or to validate a current observation using historical control data. Classically, this is done by the application of historical control limits (HCL) graphically displayed in control charts. In many applications, HCL are applied to count data, e.g. the number of revertant colonies (Ames assay) or the number of relapses per multiple sclerosis patient. Count data may be overdispersed, can be heavily right-skewed and clusters may differ in cluster size or other baseline quantities (e.g. number of petri dishes per control group or different length of monitoring times per patient). Based on the quasi-Poisson assumption or the negative-binomial distribution, we propose prediction intervals for overdispersed count data to be used as HCL. Variable baseline quantities are accounted for by offsets. Furthermore, we provide a bootstrap calibration algorithm that accounts for the skewed distribution and achieves equal tail probabilities. Comprehensive Monte-Carlo simulations assessing the coverage probabilities of eight different methods for HCL calculation reveal, that the bootstrap calibrated prediction intervals control the type-1-error best. Heuristics traditionally used in control charts (e.g. the limits in Sheward c- or u-charts or the mean plus minus 2 SD) fail to control a pre-specified coverage probability. The application of HCL is demonstrated based on data from the Ames assay and for numbers of relapses of multiple sclerosis patients. The proposed prediction intervals and the algorithm for bootstrap calibration are publicly available via the R package predint.

Sunan He, Yuxiang Nie, Hongmei Wang et al.

Generalist foundation models (GFMs) are renowned for their exceptional capability and flexibility in effectively generalizing across diverse tasks and modalities. In the field of medicine, while GFMs exhibit superior generalizability based on their extensive intrinsic knowledge as well as proficiency in instruction following and in-context learning, specialist models excel in precision due to their domain knowledge. In this work, for the first time, we explore the synergy between the GFM and specialist models, to enable precise medical image analysis on a broader scope. Specifically, we propose a cooperative framework, Generalist-Specialist Collaboration (GSCo), which consists of two stages, namely the construction of GFM and specialists, and collaborative inference on downstream tasks. In the construction stage, we develop MedDr, the largest open-source GFM tailored for medicine, showcasing exceptional instruction-following and in-context learning capabilities. Meanwhile, a series of lightweight specialists are crafted for downstream tasks with low computational cost. In the collaborative inference stage, we introduce two cooperative mechanisms, Mixture-of-Expert Diagnosis and Retrieval-Augmented Diagnosis, to harvest the generalist's in-context learning abilities alongside the specialists' domain expertise. For a comprehensive evaluation, we curate a large-scale benchmark featuring 28 datasets and about 250,000 images. Extensive results demonstrate that MedDr consistently outperforms state-of-the-art GFMs on downstream datasets. Furthermore, GSCo exceeds both GFMs and specialists across all out-of-domain disease diagnosis datasets. These findings indicate a significant paradigm shift in the application of GFMs, transitioning from separate models for specific tasks to a collaborative approach between GFMs and specialists, thereby advancing the frontiers of generalizable AI in medicine.

Miia Jansson, Maria Kääriäinen, Gillian Vesty et al.

Digital counselling may improve patients’ health outcomes, when eHealth solutions are accessible and tailored to the patients’ needs, which is especially important for people with chronic and long-term conditions such as knee osteoarthritis. This study aims to identify patients’ eHealth needs to improve the quality of digital counselling in a primary care management of symptomatic knee osteoarthritis. A qualitative study was used to collect patients’ eHealth needs through semi-structured interviews in a single outpatient clinic in Finland between August 2020 and November 2020. The data was analyzed using both deductive and inductive content analysis approaches. The study was reported in accordance with the Consolidated Criteria for Reporting Qualitative research checklist to improve the transparency of the study. Analysis of the data revealed five main categories to be considered when implementing digital counselling in patients with symptomatic knee osteoarthritis: background factors (functional impairments, health literacy, digital literacy, cost-related access barriers), resources (digital methods and materials), sufficiency (knee osteoarthritis-related knowledge and skills), implementation (simplicity, trust, patient-centeredness), and benefits (self-care capabilities, confidence). According to our findings, both health and digital literacy seems to be important contributors to the adoption of digital counselling in a primary care management of symptomatic knee osteoarthritis. New eHealth solutions should not replace the first visit in the outpatient clinic. Instead, the use of eHealth solutions should be based on the first visit, during which a trusting relationship between patients and healthcare providers is established. In future, the level of health and digital literacy in patients with symptomatic knee osteoarthritis should be taken account.

Richard Huan Xu, Bin Ma, Huanping Xin et al.

Abstract Objectives Spinal muscular atrophy (SMA) is a rare monogenic neuromuscular disorder caused by loss of function mutations. Measuring health-related quality of life to support economic evaluations in this population is encouraged. However, empirical evidence on the performance of preference-based measures (PBMs) in individuals with SMA is limited. This study aimed to assess the psychometric properties of the EQ-5D-5L and the Patient-Reported Outcomes Measure Information System Preference measure (PROPr) in individuals with SMA. Methods The data used in this study were obtained via a web-based, cross-sectional survey. All participants completed the self-reporting EQ-5D-5L and PROMIS-29 questionnaires. Information about their socioeconomic and health status was also obtained. Ceiling and floor effects, convergent and divergent validity, known-group validity, and the agreement between the two measures were assessed. Results Strong ceiling and floor effects were observed for four dimensions of the EQ-5D-5L and three subscales, including pain intensity, pain interference, and physical function, of the PROMIS-29. All hypothesized associations between EQ-5D-5L/PROMIS-29 and other neuromuscular questions were confirmed, supporting good convergent validity. Moreover, both EQ-5D-5L and PROPr scores differentiated between impaired functional groups, demonstrating good discriminative ability. Poor agreement between the EQ-5D-5L and PROPr utility scores was observed. Conclusions The EQ-5D-5L and PROPr both appear to be valid PBMs for individuals with SMA. However, PROPr yielded considerably lower utility scores than EQ-5D-5L and their agreement was poor. Therefore, these two PBMs may not be used interchangeably in economic evaluations of SMA-related interventions.

Ange Lou, Shuyue Guan, Murray Loew

Segmenting medical images accurately and reliably is important for disease diagnosis and treatment. It is a challenging task because of the wide variety of objects' sizes, shapes, and scanning modalities. Recently, many convolutional neural networks (CNN) have been designed for segmentation tasks and achieved great success. Few studies, however, have fully considered the sizes of objects, and thus most demonstrate poor performance for small objects segmentation. This can have a significant impact on the early detection of diseases. This paper proposes a Context Axial Reverse Attention Network (CaraNet) to improve the segmentation performance on small objects compared with several recent state-of-the-art models. CaraNet applies axial reserve attention (ARA) and channel-wise feature pyramid (CFP) module to dig feature information of small medical object. And we evaluate our model by six different measurement metrics. We test our CaraNet on brain tumor (BraTS 2018) and polyp (Kvasir-SEG, CVC-ColonDB, CVC-ClinicDB, CVC-300, and ETIS-LaribPolypDB) segmentation datasets. Our CaraNet achieves the top-rank mean Dice segmentation accuracy, and results show a distinct advantage of CaraNet in the segmentation of small medical objects.

Junyu Chen, Yihao Liu, Shuwen Wei et al.

Deep learning technologies have dramatically reshaped the field of medical image registration over the past decade. The initial developments, such as regression-based and U-Net-based networks, established the foundation for deep learning in image registration. Subsequent progress has been made in various aspects of deep learning-based registration, including similarity measures, deformation regularizations, network architectures, and uncertainty estimation. These advancements have not only enriched the field of image registration but have also facilitated its application in a wide range of tasks, including atlas construction, multi-atlas segmentation, motion estimation, and 2D-3D registration. In this paper, we present a comprehensive overview of the most recent advancements in deep learning-based image registration. We begin with a concise introduction to the core concepts of deep learning-based image registration. Then, we delve into innovative network architectures, loss functions specific to registration, and methods for estimating registration uncertainty. Additionally, this paper explores appropriate evaluation metrics for assessing the performance of deep learning models in registration tasks. Finally, we highlight the practical applications of these novel techniques in medical imaging and discuss the future prospects of deep learning-based image registration.