Semi-automatic Tool Research Articles

System-wide analysis of qualitative hospital incident data: Feasibility of semi-automated content analysis to uncover insights.

Advances in technology have increased the ease of reporting hospital incidents, resulting in large amounts of qualitative descriptive data. Health services have little experience analysing these data at scale to incorporate into routine reporting. We aimed to explore the feasibility of applying a semi-automated content analysis (SACA) tool (Leximancer™) to qualitative descriptions of system-wide hospital incidents to provide insights into safety issues at all health service levels. Data from 1245 incidents reported across a network of hospitals in Australia were analysed using the SACA tool. Summaries were generated using a variety of techniques, including inductive and deductive approaches to extract key concepts in the data. The analysis was feasible and provided an actionable summary of the types of incidents reported in the data; the visual interface allowed users to explore the underlying text for a deeper understanding. Deductive analysis was utilised to explore specific areas of interest, and stratified analysis revealed more detailed concepts. The SACA tool was more efficient than manual processes; however, due to the context present in the incident descriptions, significant time, reading and subject matter expertise is still required to refine the analysis. Semi-automated tools provide an opportunity for improving patient safety culture and practices by providing rapid content analysis of vast datasets that can be customised for specific organisational contexts and deployed at scale. Further research is required to assess usefulness with system users. Qualitative data abound and system-wide analysis is essential to creating actionable insights.

Mobile app review analysis for crowdsourcing of software requirements: a mapping study of automated and semi-automated tools

Mobile app reviews are valuable for gaining user feedback on features, usability, and areas for improvement. Analyzing these reviews manually is difficult due to volume and structure, leading to the need for automated techniques. This mapping study categorizes existing approaches for automated and semi-automated tools by analyzing 180 primary studies. Techniques include topic modeling, collocation finding, association rule-based, aspect-based sentiment analysis, frequency-based, word vector-based, and hybrid approaches. The study compares various tools for analyzing mobile app reviews based on performance, scalability, and user-friendliness. Tools like KEFE, MERIT, DIVER, SAFER, SIRA, T-FEX, RE-BERT, and AOBTM outperformed baseline tools like IDEA and SAFE in identifying emerging issues and extracting relevant information. The study also discusses limitations such as manual intervention, linguistic complexities, scalability issues, and interpretability challenges in incorporating user feedback. Overall, this mapping study outlines the current state of feature extraction from app reviews, suggesting future research and innovation opportunities for extracting software requirements from mobile app reviews, thereby improving mobile app development.

A Comparative Crash-Test of Manual and Semi-Automated Methods for Detecting Complex Submarine Morphologies

Multibeam echosounders provide ideal data for the semi-automated seabed feature extraction and accurate morphometric measurements. In this study, bathymetric and raw backscatter data were initially used to manually delimit the reef morphologies found in an insular semi-enclosed gulf in the northern Aegean Sea (Gera Gulf, Lesvos Island, Greece). The complexity of this environment makes it an ideal area to “crash test” (test to the limit) and compare the results of the delineation methods. A large number of (more than 7000) small but prominent reefs were detected, which made manual mapping extremely time-consuming. Three semi-automated tools were also employed to map the reefs: the Benthic Terrain Modeler (BTM), Confined Morphologies Mapping (CoMMa), and eCognition Multiresolution Segmentation. BTM did not function properly with irregular reef footprints, but by modifying both the bathymetry and slope, the outcome was improved, producing accurate results that appeared to exceed the accuracy of manual mapping. CoMMa, a new GIS morphometric toolbox, was a “one-stop shop” that, besides generating satisfactory reef delineation results (i.e., detecting the same total reef area as the manual method), was also used to extract the morphometric characteristics of the polygons resulting from all the methods. Lastly, the Multiresolution Segmentation also gave satisfactory results with the highest precision. To compare the final maps with the distribution of the reefs, mapcurves were created to estimate the goodness-of-fit (GOF) with the Precision, Recall, and F1 Scores producing values higher than 0.78, suggesting a good detection accuracy for the semi-automated methods. The analysis reveals that the semi-automated methods provided more efficient results in comparison with the time-consuming manual mapping. Overall, for this case study, the modification of the bathymetry and slope enabled the results’ accuracy to be further enhanced. This study asserts that the use of semi-automated mapping is an effective method for delineating the geomorphometry of intricate relief and serves as a powerful tool for habitat mapping and decision-making.

Diabetes worsens the adverse association between myocardial diffuse fibrosis and left ventricular strain

Abstract Background Heart failure is a common cardiovascular complication to diabetes. Increased fibrosis of the myocardium has been suggested to be a mechanism, however, no convincing clinical evidence has so far been presented. A better understanding of how myocardial tissue characteristics influence functional aberration in diabetes could aid diagnosis, risk assessment and guide treatment. Native T1 is a CMR marker of diffuse fibrosis with diagnostic and prognostic value (1,2), and feature tracking (CMR-FT) strain is a sensitive measure of myocardial systolic function. In this study, we evaluated the association of CMR markers of fibrosis with markers of systolic function in individuals with and without diabetes. Purpose To understand the relationship between native T1 and myocardial strain and how diabetes may affect this. Methods Participants from the UK Biobank population imaging study were included. Left ventricular global longitudinal strain (GLS) was derived using a semi-automated batch processing software tool. Native T1 was derived from segmentation of the entire short axis slice using an automated quality control tool, excluding scans with predicted dice score of <0.7. Multivariable linear models were used with the strain metrics as the outcome variable and native T1 and diabetes as the exposure variables. These models were adjusted for age, sex, BMI, ethnicity, smoking and alcohol status, prevalent hypertension hypercholesterolaemia and coronary artery disease. The effect of interaction between diabetes and native T1 on GLS was examined by introducing an interaction term. Results In total 41,533 participants were included, of which 1957 had diabetes – see Table 1 for baseline characteristics. The average GLS measured for those with diabetes was -17% and was -18% for those without. The average native T1 in those with diabetes was 935ms and 933ms for those without on the background of greater prevalence of cardiovascular risk factors in those with diabetes. In model adjusted for all relevant co-variates, individuals with diabetes had worse GLS measured compared to those without diabetes (β-coefficient: 0.51, p value <0.001). Increasing fibrosis (native T1) was associated with worsening GLS (β-coefficient: 0.005, p value <0.001) in all participants. However, in those with diabetes, increasing native T1 was associated with a greater decline in GLS compared to those without diabetes (interaction term β-coefficient: 0.005, p value <0.001), with a relative impairment seen within normal ranges for native T1 - see Figure 1. Conclusion(s) This is the largest study of its kind to demonstrate a dose dependent relationship between increasing native T1, representing potential myocardial fibrosis, and myocardial strain, representing cardiac dysfunction. Our findings indicate that increased fibrosis alone does not explain the higher risk of HF in diabetes, other mechanisms, perhaps intracellular, might be at play.

Software-assisted structured reporting and semi-automated TNM classification for NSCLC staging in a multicenter proof of concept study

ObjectivesIn this multi-center study, we proposed a structured reporting (SR) framework for non-small cell lung cancer (NSCLC) and developed a software-assisted tool to automatically translate image-based findings and annotations into TNM classifications. The aim of this study was to validate the software-assisted SR tool for NSCLC, assess its potential clinical impact in a proof-of-concept study, and evaluate current reporting standards in participating institutions.MethodsA framework for SR and staging of NSCLC was developed in a multi-center collaboration. SR annotations and descriptions were used to generate semi-automated TNM classification. The SR and TNM classification tools were evaluated by nine radiologists on n = 20 representative [18F]FDG PET/CT studies and compared to the free text reporting (FTR) strategy. Results were compared to a multidisciplinary team reference using a generalized linear mixed model (GLMM). Additionally, participants were surveyed on their experience with SR and TNM classification.ResultsOverall, GLMM analysis revealed that readers using SR were 1.707 (CI: 1.137–2.585) times more likely to correctly classify TNM status compared to FTR strategy (p = 0.01) resulting in increased overall TNM correctness in 71.9% (128/178) of cases compared to 62.8% (113/180) FTR. The primary source of variation in classification accuracy was explained by case complexity. Participants rated the potential impact of SR and semi-automated TNM classification as positive across all categories with improved scores after template validation.ConclusionThis multi-center study yielded an effective software-assisted SR framework for NSCLC. The SR and semi-automated classification tool improved TNM classification and were perceived as valuable.Critical relevance statementSoftware-assisted SR provides robust input for semi-automated rule-based TNM classification in non-small-cell lung carcinoma (NSCLC), improves TNM correctness compared to FTR, and was perceived as valuable by radiology physicians.Key PointsSR and TNM classification are underutilized across participating centers for NSCLC staging.Software-assisted SR has emerged as a promising strategy for oncologic assessment.Software-assisted SR facilitates semi-automated TNM classification with improved staging accuracy compared to free-text reports in NSCLC.Graphical

A complementary approach for detecting biological signals through a semi-automated feature selection tool.

Untargeted metabolomics is often used in studies that aim to trace the metabolic profile in a broad context, with the data-dependent acquisition (DDA) mode being the most commonly used method. However, this approach has the limitation that not all detected ions are fragmented in the data acquisition process, in addition to the lack of specificity regarding the process of fragmentation of biological signals. The present work aims to extend the detection of biological signals and contribute to overcoming the fragmentation limits of the DDA mode with a dynamic procedure that combines experimental and in silico approaches. Metabolomic analysis was performed on three different species of actinomycetes using liquid chromatography coupled with mass spectrometry. The data obtained were preprocessed by the MZmine software and processed by the custom package RegFilter. RegFilter allowed the coverage of the entire chromatographic run and the selection of precursor ions for fragmentation that were previously missed in DDA mode. Most of the ions selected by the tool could be annotated through three levels of annotation, presenting biologically relevant candidates. In addition, the tool offers the possibility of creating local spectral libraries curated according to the user's interests. Thus, the adoption of a dynamic analysis flow using RegFilter allowed for detection optimization and curation of potential biological signals, previously absent in the DDA mode, being a good complementary approach to the current mode of data acquisition. In addition, this workflow enables the creation and search of in-house tailored custom libraries.

Accelerated muscle mass estimation from CT images through transfer learning

BackgroundThe cost of labeling to collect training data sets using deep learning is especially high in medical applications compared to other fields. Furthermore, due to variances in images depending on the computed tomography (CT) devices, a deep learning based segmentation model trained with a certain device often does not work with images from a different device.MethodsIn this study, we propose an efficient learning strategy for deep learning models in medical image segmentation. We aim to overcome the difficulties of segmentation in CT images by training a VNet segmentation model which enables rapid labeling of organs in CT images with the model obtained by transfer learning using a small number of manually labeled images, called SEED images. We established a process for generating SEED images and conducting transfer learning a model. We evaluate the performance of various segmentation models such as vanilla UNet, UNETR, Swin-UNETR and VNet. Furthermore, assuming a scenario that a model is repeatedly trained with CT images collected from multiple devices, in which is catastrophic forgetting often occurs, we examine if the performance of our model degrades.ResultsWe show that transfer learning can train a model that does a good job of segmenting muscles with a small number of images. In addition, it was confirmed that VNet shows better performance when comparing the performance of existing semi-automated segmentation tools and other deep learning networks to muscle and liver segmentation tasks. Additionally, we confirmed that VNet is the most robust model to deal with catastrophic forgetting problems.ConclusionIn the 2D CT image segmentation task, we confirmed that the CNN-based network shows better performance than the existing semi-automatic segmentation tool or latest transformer-based networks.

CyCadas: accelerating interactive annotation and analysis of clustered cytometry data.

Single cell profiling by cytometry has emerged as a key technology in biology, immunology and clinical-translational medicine. The correct annotation, which refers to the identification of clusters as specific cell populations based on their marker expression, of clustered high-dimensional cytometry data, is a critical step of the analysis. Its accuracy determines the correct interpretation of the biological data. Despite the progress in various clustering algorithms, the annotation of clustered data still remains a manual, time consuming and error-prone task. We developed a user-friendly cluster annotation and differential abundance detection tool that can be applied on data generated with Self Organizing Map clustering algorithms, thus simplifying the annotation process of datasets that consist of hundreds or thousands of clusters. We present Cytometry Cluster Annotation and Differential Abundance Suite (CyCadas), a semi-automated software tool that facilitates cluster annotation in cytometry data by offering both visual and computational guidance. CyCadas addresses the critical need for efficient and accurate annotation of high-resolution clustered cytometry data, significantly reducing the time needed to perform the analysis compared to both manual gating approaches and manual annotation of clustered data. The tool features a user-friendly interface, visual tools enabling data exploration and automated threshold estimation to separate negative and positive marker expression. It facilitates the definition and annotation of cell phenotypes among multiple clusters in a tree-based data structure. Finally, it calculates the abundance of various cell populations across the conditions with statistical interpretation. It is an ideal resource for researchers aiming to streamline their cytometry workflow. CyCadas is available as open source at: https://github.com/DII-LIH-Luxembourg/cycadas.

An in silico design method of a peptide bioreceptor for cortisol using molecular modelling techniques

Cortisol is established as a reliable biomarker for stress prompting intensified research in developing wearable sensors to detect it via eccrine sweat. Since cortisol is present in sweat in trace quantities, typically 8–140 ng/mL, developing such biosensors necessitates the design of bioreceptors with appropriate sensitivity and selectivity. In this work, we present a systematic biomimetic methodology and a semi-automated high-throughput screening tool which enables rapid selection of bioreceptors as compared to ab initio design of peptides via computational peptidology. Candidate proteins from databases are selected via molecular docking and ranked according to their binding affinities by conducting automated AutoDock Vina scoring simulations. These candidate proteins are then validated via full atomistic steered molecular dynamics computations including umbrella sampling to estimate the potential of mean force using GROMACS version 2022.6. These explicit molecular dynamic calculations are carried out in an eccrine sweat environment taking into consideration the protein dynamics and solvent effects. Subsequently, we present a candidate baseline peptide bioreceptor selected as a contiguous sequence of amino acids from the selected protein binding pocket favourably interacting with the target ligand (i.e., cortisol) from the active binding site of the proteins and maintaining its tertiary structure. A unique cysteine residue introduced at the N-terminus allows orientation-specific surface immobilization of the peptide onto the gold electrodes and to ensure exposure of the binding site. Comparative binding affinity simulations of this peptide with the target ligand along with commonly interfering species e.g., progesterone, testosterone and glucose are also presented to demonstrate the validity of this proposed peptide as a candidate baseline bioreceptor for future cortisol biosensor development.

Predicting image quality of forensic footwear impressions

The ability of a footwear examiner to confidently discern features of importance in a forensic examination is directly related to impression quality. As a result, quality directly impacts the strength an examiner can ascribe to any opinion of source attribution. Despite the importance of image quality during both the analysis and comparison phases of an examination, there is limited research on the estimation, variation, and prediction of footwear impression quality. In response, this study aims to develop a methodology for evaluating footwear impression quality by regressing image features against subjective judgments of quality. Using a dataset of more than 450 impressions evaluated by more than 40 participants, estimates of intra- and inter-rater consistency were computed. After identifying reliable raters, matrix completion was performed, thereby permitting data imputation. This was based on an approximate 90:10 split between training and testing data, ultimately resulting in 6,000 quality predictions, which exhibited a mean agreement with ground truth > 95%. Each image in the dataset was then reduced to a 10-dimensional feature vector describing image attributes such as complexity, contrast, sharpness, and noise. These features were used to train multiple ordinal and multinomial regression models aimed at predicting image quality. Model comparison resulted in an overall optimal accuracy of approximately 75% when comparing predicted/modeled quality against the ground truth of subjective rater opinions. The resulting semi-automated, reference-free and numerical prediction tool exhibits reasonable success for impression quality prediction when presented with the types of challenging and domain-specific images often encountered in forensic footwear comparisons. The outcomes and challenges associated with this investigation provide a foundation upon which future studies of quality can be modified and built, with the downstream goal of an increased understanding of how quality impacts weight of evidence.

Distribution of MRI-derived T2 values as a biomarker for in vivo rapid screening of phenotype severity in mdx mice

Background The pathology in Duchenne muscular dystrophy (DMD) is characterized by degenerating muscle fibers, inflammation, fibro-fatty infiltrate, and edema, and these pathological processes replace normal healthy muscle tissue. The mdx mouse model is one of the most commonly used preclinical models to study DMD. Mounting evidence has emerged illustrating that muscle disease progression varies considerably in mdx mice, with inter-animal differences as well as intra-muscular differences in pathology in individual mdx mice. This variation is important to consider when conducting assessments of drug efficacy and in longitudinal studies. We developed a magnetic resonance imaging (MRI) segmentation and analysis pipeline to rapidly and non-invasively measure the severity of muscle disease in mdx mice. Methods Wildtype and mdx mice were imaged with MRI and T2 maps were obtained axially across the hindlimbs. A neural network was trained to rapidly and semi-automatically segment the muscle tissue, and the distribution of resulting T2 values was analyzed. Interdecile range and Pearson Skew were identified as biomarkers to quickly and accurately estimate muscle disease severity in mice. Results The semiautomated segmentation tool reduced image processing time approximately tenfold. Measures of Pearson skew and interdecile range based on that segmentation were repeatable and reflected muscle disease severity in healthy wildtype and diseased mdx mice based on both qualitative observation of images and correlation with Evans blue dye uptake. Conclusion Use of this rapid, non-invasive, semi-automated MR image segmentation and analysis pipeline has the potential to transform preclinical studies, allowing for pre-screening of dystrophic mice prior to study enrollment to ensure more uniform muscle disease pathology across treatment groups, improving study outcomes.

Improving Reproducibility of Volumetric Evaluation Using Computed Tomography in Pediatric Patients with Congenital Heart Disease.

The volumetric data obtained from the cardiac CT scan of congenital heart disease patients is important for defining patient's status and making decision for proper management. The objective of this study is to evaluate the intra-observer, inter-observer, and interstudy reproducibility of left ventricular (LV) and right ventricular (RV) or functional single-ventricle (FSV) volume. And compared those between manual and using semi-automated segmentation tool. Total of 127 patients (56 female, 71 male; mean age 82.1months) underwent pediatric protocol cardiac CT from January 2020 to December 2022. The volumetric data including both end-systolic and -diastolic volume and calculated EF were derived from both conventional semiautomatic region growing algorithms (CM, TeraRecon, TeraRecon, Inc., San Mateo, CA, USA) and deep learning-based annotation program (DLS, Medilabel, Ingradient, Inc., Seoul, Republic of Korea) by three readers, who have different background knowledge or experience of radiology or image extraction before. The reproducibility was compared using intra- and inter-observer agreements. And the usability was measured using time for reconstruction and number of tests that were reconfigured before the reconfiguration time was reduced to less than 5min. Inter- and intra-observer agreements showed better agreements degrees in DLS than CM in all analyzers. The time used for reconstruction showed significantly shorter in DLS compared with CM. And significantly small numbers of tests before the reconfiguration is needed in DLS than CM. Deep learning-based annotation program can be more accurate way for measurement of volumetric data for congenital heart disease patients with better reproducibility than conventional method.

Semi-automated vertex placement for lattice radiotherapy and dosimetric verification using 3D polymer gel dosimetry.

To evaluate the feasibility of an open-source, semi-automated, and reproducible vertex placement tool to improve the efficiency of lattice radiotherapy (LRT) planning. We used polymer gel dosimetry with a Cone Beam CT (CBCT) readout to commission this LRT technique. We generated a volumetric modulated arc therapy (VMAT)-based LRT plan on a 2 L NIPAM polymer gel dosimeter using our Eclipse Acuros version 15.6 AcurosXB beam model, and also recalculated the plan with a pre-clinical Acuros v18.0 dose calculation algorithm with the enhanced leaf modelling (ELM). With the assistance of the MAAS-SFRThelper software, a lattice vertex diameter of 1.5cm and center-to-center spacing of 3cm were used to place the spheres in a hexagonal, closed packed structure. The verification plan included four gantry arcs with 15°, 345°, 75°, 105° collimator angles. The spheres were prescribed 20Gy to 50% of their combined volume. The 6 MV Flattening Filter Free beam energy was used to deliver the verification plan. The dosimetric accuracy of the LRT delivery was evaluated with 1D dose profiles, 2D isodose maps, and a 3D global gamma analysis. Qualitative comparisons between the 1D dose profiles of the Eclipse plan and measured gel showed good consistency at the prescription dose mark. The average diameter measured 13.3±0.2mm (gel for v15.6), 12.6mm (v15.6 plan), 13.1±0.2mm (gel for v18.0), and 12.3mm (v18.0 plan). 3D gamma analysis showed that all gamma pass percent were>95% except at 1% and 2% at the 1mm distance to agreement criteria. This study presents a novel application of gel dosimetry in verifying the dosimetric accuracy of LRT, achieving excellent 3D gamma results. The treatment planning was facilitated by publicly available software that automatically placed the vertices for consistency and efficiency.

BuildAMol: a versatile Python toolkit for fragment-based molecular design

In recent years computational methods for molecular modeling have become a prime focus of computational biology and cheminformatics. Many dedicated systems exist for modeling specific classes of molecules such as proteins or small drug-like ligands. These are often heavily tailored toward the automated generation of molecular structures based on some meta-input by the user and are not intended for expert-driven structure assembly. Dedicated manual or semi-automated assembly software tools exist for a variety of molecule classes but are limited in the scope of structures they can produce. In this work we present BuildAMol, a highly flexible and extendable, general-purpose fragment-based molecular assembly toolkit. Written in Python and featuring a well-documented, user-friendly API, BuildAMol empowers researchers with a framework for detailed manual or semi-automated construction of diverse molecular models. Unlike specialized software, BuildAMol caters to a broad range of applications. We demonstrate its versatility across various use cases, encompassing generating metal complexes or the modeling of dendrimers or integrated into a drug discovery pipeline. By providing a robust foundation for expert-driven model building, BuildAMol holds promise as a valuable tool for the continuous integration and advancement of powerful deep learning techniques.Scientific contributionBuildAMol introduces a cutting-edge framework for molecular modeling that seamlessly blends versatility with user-friendly accessibility. This innovative toolkit integrates modeling, modification, optimization, and visualization functions within a unified API, and facilitates collaboration with other cheminformatics libraries. BuildAMol, with its shallow learning curve, serves as a versatile tool for various molecular applications while also laying the groundwork for the development of specialized software tools, contributing to the progress of molecular research and innovation.

One Size Does Not Fit All: Micro-, Meso-, and Macrobleeds in Cerebral Amyloid Angiopathy

Introduction: MRI rating criteria for small vessel disease markers include definitions for microbleeds and macrobleeds but do not account for small (<10 mm) hemorrhages with a cystic cavity and/or irregular shape. Such hemorrhages, however, are often present in patients with cerebral amyloid angiopathy (CAA). In this study, we aimed to investigate the frequency, diameter, and volume distribution of these hemorrhages (which we called mesobleeds) in patients with CAA. Methods: We selected participants with Dutch-type hereditary CAA (D-CAA) and sporadic CAA (sCAA) and scored microbleeds, mesobleeds, and macrobleeds on 3T susceptibility-weighted images MRI. Hemorrhage diameter and volume were calculated in a subset of participants using a semi-automatic tool; their distribution was evaluated on a logarithmic scale. Results: We included 25 participants with D-CAA (mean age 56 years) and 25 with sCAA (mean age 73 years). In total, 11,007 microbleeds, 602 mesobleeds, and 195 macrobleeds were observed. Eighty-two percent of participants had ≥1 mesobleed. Hemorrhage diameter and volume were calculated in four participants with 272 microbleeds (median diameter 1.52 mm, volume 0.004 mL), 84 mesobleeds (median diameter 5.61 mm, volume 0.06 mL), and 37 macrobleeds (median diameter 19.58 mm, volume 1.33 mL). Mesobleed diameter and volume were larger than microbleeds (optimal cut-off 0.02 mL) but showed overlap with macrobleeds. Conclusion: Hemorrhages <10 mm with an irregular shape and/or cystic cavity are frequently found in participants with CAA and have a distinct diameter and volume distribution. We propose to name these hemorrhage mesobleeds and to rate them separately from micro- and macrobleeds. Future research is necessary to investigate their pathophysiology and prognostic value.

Efficiency and Workload Reduction of Semi-automated Citation Screening Software for Creating Clinical Practice Guidelines: A Prospective Observational Study.

We evaluated the applicability of automated citation screening in developing clinical practice guidelines. We prospectively compared the efficiency of citation screening between the conventional (Rayyan) and semi-automated (ASReview software) methods. We searched the literature for five clinical questions (CQs) in the development of the Japanese Clinical Practice Guidelines for the Management of Sepsis and Septic Shock. Objective measurements of the time required to complete citation screening were recorded. Following the first screening round, in the primary analysis, the sensitivity, specificity, positive predictive value, and overall screening time were calculated for both procedures using the semi-automated tool as index and the results of the conventional method as standard reference. In the secondary analysis, the same parameters were compared between the two procedures using the final list of included studies after the second screening session as standard reference. Among the five CQs after the first screening session, the highest and lowest sensitivity, specificity, and positive predictive values were 0.241 and 0.795; 0.991 and 1.000; and 0.482 and 0.929, respectively. In the secondary analysis, the highest sensitivity and specificity in the semi-automated citation screening were 1.000 and 0.997, respectively. The overall screening time per 100 studies was significantly shorter with semi-automated than with conventional citation screening. The potential advantages of the semi-automated method (shorter screening time and higher discriminatory rate for the final list of studies) warrant further validation.

A Simplified Framework to Integrate Databases with Building Information Modeling for Building Energy Assessment in Multi-Climate Zones

BIM models are seldom used for the energy certification of buildings. This paper discusses the advantages of linking two important fields: building information modeling (BIM) and building environmental assessment methods (BEAM), presented as a rating system and a proposal for the Chilean context. The state of the art in both fields around the world is discussed, with an in-depth examination of current BIM software and related applications, followed by a discussion about previous research on integrating them. A lack of interoperability and data losses between BIM and BEM were found. A new tool is presented that addresses these challenges to ensure accurate rating system data, and this new framework is based on database exchange and takes crucial information from BIM to BEAM platforms. The development of the method includes BIM programming (API), database links, and spreadsheets for a Chilean building energy certification through a new tool, also applicable to multiclimactic zones. This new semi-automatic tool allows architects to model their design in a BIM platform and use this information as input for the energy certification process. The potential and risks of this method are discussed. Several improvements and enhancements of the energy certification process were found when incorporating this new framework in comparison to current methodologies.

Undisturbed culture: a clinical examination of this culture strategy on embryo in vitro development and clinical outcomes

To compare the effect of a fully undisturbed culture strategy over a sequential one on embryo in vitro development and clinical outcomes in ICSI cycles. Retrospective cohort study. This study included 4,564 ICSI cycles performed over 5 years, including autologous and oocyte donation treatments with extended embryo culture until blastocyst in one of the two defined culture strategies. Embryo cohorts were cultured in one of two culture systems: a fully undisturbed culture, including an incubator with integrated time-lapse technology, one-step culture medium and embryo selection assisted by semi-automatic tools based on embryo morphokinetics, or a sequential culture, using a conventional benchtop incubator, sequential media and traditional morphological evaluation under optical microscope. The effect of the culture strategies over embryo development and clinical outcomes was quantified by generalized estimated equations, controlling for possible confounders through the inverse probability of treatment weighting method. Weighted odds ratios (OR) and 95% confidence intervals (CI) for live birth rate after fresh single embryo transfer and the cumulative live birth rate. Additionally, blastocyst development and morphology and other intermediate outcomes were also assessed. A significant positive association was found between the employment of undisturbed embryo culture and higher live birth rate in the first embryo transfer in both autologous (OR=1.617, 95%CI: 1.074-2.435) and oocyte donation cycles (OR=1.316, 95%CI: 1.036-1.672). Cumulative live birth rate after one year follow-up was also positively associated with the undisturbed culture strategy in oocyte donation cycles (OR=1.5, 95%CI: 1.179-1.909), but not in autologous cycles (OR=1.051, 95%CI: 0.777-1.423). Similarly, blastocyst rate, good morphology blastocyst rate and utilisation rate were positively associated with the employment of undisturbed culture in oocyte donation cycles, but not in autologous cycles. These findings imply that a culture system combining integrated time-lapse incubators with a one-step culture medium, may enhance the success rates of patients undergoing ICSI treatment by increasing the production of higher-quality blastocysts and improving embryo selection while streamlining laboratory procedures and workflow.

Could this be next for corpus linguistics? Methods of semi-automatic data annotation with contextualized word embeddings

Abstract This paper explores how linguistic data annotation can be made (semi-)automatic by means of machine learning. More specifically, we focus on the use of “contextualized word embeddings” (i.e. vectorized representations of the meaning of word tokens based on the sentential context in which they appear) extracted by large language models (LLMs). In three example case studies, we assess how the contextualized embeddings generated by LLMs can be combined with different machine learning approaches to serve as a flexible, adaptable semi-automated data annotation tool for corpus linguists. Subsequently, to evaluate which approach is most reliable across the different case studies, we use a Bayesian framework for model comparison, which estimates the probability that the performance of a given classification approach is stronger than that of an alternative approach. Our results indicate that combining contextualized word embeddings with metric fine-tuning yield highly accurate automatic annotations.

Longitudinal Lesion Expansion in Chronic Traumatic Brain Injury.

Traumatic brain injury (TBI) is a risk factor for neurodegeneration and cognitive decline, yet the underlying pathophysiologic mechanisms are incompletely understood. This gap in knowledge is in part related to the lack of analytic methods to account for cortical lesions in prior neuroimaging studies. The objective of this study was to develop a lesion detection tool and apply it to an investigation of longitudinal changes in brain structure among individuals with chronic TBI. We identified 24 individuals with chronic moderate-to-severe TBI enrolled in the Late Effects of TBI (LETBI) study who had cortical lesions detected by T1-weighted MRI at two time points. Initial MRI scans were performed more than 1-year post-injury and follow-up scans were performed 3.1 (IQR=1.7) years later. We leveraged FreeSurfer parcellations of T1-weighted MRI volumes and a recently developed super-resolution technique, SynthSR, to identify cortical lesions in this longitudinal dataset. Trained raters received the data in a randomized order and manually corrected the automated lesion segmentation, yielding a final lesion mask for each scan at each timepoint. Lesion volume significantly increased between the two time points with a median volume change of 3.2 (IQR=5.9) mL (p<0.001), and the increases significantly exceeded the possible variance in lesion volume changes due to manual tracing errors (p < 0.001). Lesion volume significantly expanded longitudinally in 23 of 24 subjects, with all FDR corrected p-values ≤ 0.02. Inter-scan duration was not associated with the magnitude of lesion growth. We also demonstrated that the semi-automated tool showed a high level of accuracy compared to "ground truth" manual lesion segmentation. Semi-automated lesion segmentation is feasible in TBI studies and creates opportunities to elucidate mechanisms of post-traumatic neurodegeneration.