Comparing Performance Metrics Research Articles

Analysis of potential microRNA biomarkers for multiple sclerosis

Multiple sclerosis (MS) is a chronic demyelinating autoimmune neurodegenerative disorder for which no specific blood biomarker is available. MicroRNAs (miRNAs) have been investigated for their diagnostic potential in MS. However, MS-associated miRNAs are rarely replicated in different MS populations, thus impeding their use in clinical testing. Here, we evaluated the fold expression of seven reported MS miRNAs associated with MS incidence and clinical characteristics in 76 MS patients and 75 healthy control plasma samples. We found miR-23a-3p to be upregulated in relapsing-remitting MS (RRMS), while miR-326 was downregulated. MiR-150-5p and -320a-3p were significantly downregulated in secondary progressive MS (SPMS) patients compared to RRMS. High disability was associated with low miR-320a-3p, whereas low BDNF levels were associated with upregulation of miR-150-5p and downregulation of miR-326 expression in the total cohort. MiR-23a-3p and miR-326 showed significant diagnostic sensitivity, specificity, and accuracy for RRMS diagnosis. In addition, miR-150-5p and miR-320a-3p had comparable significant diagnostic test performance metrics distinguishing SPMS from RRMS. Therefore, there is potential for including miR-23a-3p and miR-326 in an RRMS diagnostic miRNA panel. Moreover, we have shown that miR-150-5p and miR-320a-3p could be novel RRMS conversion to SPMS biomarkers. The use of these miRNAs in MS diagnosis and prognosis warrants further investigation.

Efficient Breast Cancer Dataset Analysis Based on Adaptive Classifiers

Many algorithms have been used to diagnose diseases, with some demonstrating good performance while others have not met expectations. Making correct decisions with the minimal possible errors is of the highest priority when diagnosing diseases. Breast cancer, being a prevalent and widespread disease, emphasizes the importance of early detection. Accurate decision-making regarding breast cancer is crucial for early treatment and achieving favorable outcomes. The percentage split evaluation approach was employed, comparing performance metrics such as precision, recall, and f1-score. Kernel Naïve Bayes achieved 100% precision in the percentage split method for breast cancer, while the Coarse Gaussian support vector machines achieved 97.2% precision in classifying breast cancer in 4-fold cross-validation.

Mind your prevalence!

Multiple metrics are used when assessing and validating the performance of quantitative structure–activity relationship (QSAR) models. In the case of binary classification, balanced accuracy is a metric to assess the global performance of such models. In contrast to accuracy, balanced accuracy does not depend on the respective prevalence of the two categories in the test set that is used to validate a QSAR classifier. As such, balanced accuracy is used to overcome the effect of imbalanced test sets on the model’s perceived accuracy. Matthews' correlation coefficient (MCC), an alternative global performance metric, is also known to mitigate the imbalance of the test set. However, in contrast to the balanced accuracy, MCC remains dependent on the respective prevalence of the predicted categories. For simplicity, the rest of this work is based on the positive prevalence. The MCC value may be underestimated at high or extremely low positive prevalence. It contributes to more challenging comparisons between experiments using test sets with different positive prevalences and may lead to incorrect interpretations. The concept of balanced metrics beyond balanced accuracy is, to the best of our knowledge, not yet described in the cheminformatic literature. Therefore, after describing the relevant literature, this manuscript will first formally define a confusion matrix, sensitivity and specificity and then present, with synthetic data, the danger of comparing performance metrics under nonconstant prevalence. Second, it will demonstrate that balanced accuracy is the performance metric accuracy calibrated to a test set with a positive prevalence of 50% (i.e., balanced test set). This concept of balanced accuracy will then be extended to the MCC after showing its dependency on the positive prevalence. Applying the same concept to any other performance metric and widening it to the concept of calibrated metrics will then be briefly discussed. We will show that, like balanced accuracy, any balanced performance metric may be expressed as a function of the well-known values of sensitivity and specificity. Finally, a tale of two MCCs will exemplify the use of this concept of balanced MCC versus MCC with four use cases using synthetic data.Scientific contributionThis work provides a formal, unified framework for understanding prevalence dependence in model validation metrics, deriving balanced metric expressions beyond balanced accuracy, and demonstrating their practical utility for common use cases. In contrast to prior literature, it introduces the derived confusion matrix to express metrics as functions of sensitivity, specificity and prevalence without needing additional coefficients. The manuscript extends the concept of balanced metrics to Matthews' correlation coefficient and other widely used performance indicators, enabling robust comparisons under prevalence shifts.

Empowering Structural Integrity: YOLO-Based Crack Detection and ML-Driven Concrete Strength Prediction of Critical Infrastructure Caused due to Mining Operation.

In this paper, YOLO algorithm is used for detecting crack by incorporating machine learning approach utilized for predicting concrete strength and possible crack in critical infrastructure caused during mining operation. Over time, those fissures become more noticeable until they finally collapse. Big data is accumulating huge amount of data by utilizing remote sensing, earth penetrating RADAR, LiDAR, GPS, and satellite image data for large-scale operations. The performance of the proposed approach is compared with other state-of-the-art algorithms, such as, YOLO V3 and YOLO V8. To investigate the performance, four performances metric like mean absolute error (MAE), root mean square error (RMSE), relative standard error (RSE) and root relative squared error (RRSE) has considered. It is found that our approach is giving superior results in term of the compared performance metrics. Moreover, non-parametric statistical testing, such as, signed test, Wilcoxon Rank test and Friedman's test has been conducted to verify the dominance of proposed approach over others.

Comparative Analysis Using Machine Learning Algorithms to Detect Parkinson’s Disease using Voice Dataset

Abstract: Parkinson's disease is a degenerative disorder that affects the control of movement, and its symptoms include tremors, stiffness, and difficulties in balance and coordination. This disease results from a decrease in dopamine-producing cells in the brain, and although there is currently no cure for Parkinson's, medication and therapeutic interventions can help alleviate the symptoms. Parkinson's disease usually begins to show its effects in middle to late adulthood. A new measure of dysphonia, Pitch Period Entropy (PPE) assesses the variability or randomness of pitch periods in speech signals, providing insights into the diversity of vocal characteristics. Higher entropy values suggest greater pitch variability in a speaker's voice. PPE is included in the dataset. The methodologies used in this project are Random Forest, Support vector classifier, Extreme gradient boosting (XGB). This comparative study includes performance metrics: Accuracy, precision, Recall, f1 score, and support. The aim is to predict Parkinson's disease and compare performance metrics using various methods.

Routing Selection Algorithm for Mobile Ad Hoc Networks Based on Neighbor Node Density.

In the process of data transmission in mobile ad hoc networks, it is essential to establish optimal routes from source nodes to destination nodes. However, as network density increases, this process is often accompanied by a significant rise in network overhead. To address this issue, the ND-AODV (neighborhood density AODV) protocol has been introduced, which reduces the probability of transmitting control information in high-density node environments to mitigate network overhead. Nevertheless, this may come at the cost of reduced routing accuracy, potentially leading to unnecessary resource wastage in certain scenarios. Furthermore, ND-AODV does not comprehensively consider the location of the receiving nodes, which limits its ability to reduce network overhead effectively. To overcome these limitations, this paper introduces a novel routing approach, known as CND-AODV (common neighborhood density AODV). In comparison to ND-AODV, CND-AODV offers a more comprehensive solution to the challenges posed by high-density network environments. It intelligently processes control information based on the special positioning of the receiving nodes, thereby significantly reducing unnecessary network overhead. Through simulation experiments comparing performance metrics such as throughput, packet delivery rate, and latency, the results clearly indicate that CND-AODV substantially decreases network overhead, enhancing network performance. Compared to ND-AODV, this innovative routing approach exhibits significant advantages. It provides a more efficient and reliable solution for ad hoc networks in high-density environments.

A Comparison of Preferentially Expressed Antigen in Melanoma Immunohistochemistry and Diagnostic Gene Expression-Profiling Assay in Challenging Melanocytic Proliferations.

Most melanocytic tumors are classified as benign or malignant based on clinical morphology, histology, and immunohistochemical (IHC) analysis. A subset of more challenging cases with ambiguous features may require further evaluation with established ancillary diagnostic molecular studies, including fluorescence in situ hybridization and/or single nucleotide polymorphism array, to increase diagnostic certainty. More recently, a diagnostic gene expression-profiling (GEP) assay and an IHC stain for the detection of PRAME (PReferentially expressed Antigen in MElanoma) have been developed. The use of PRAME IHC has been validated in cases of unequivocal and ambiguous melanocytic proliferations via comparing results with fluorescence in situ hybridization and/or single nucleotide polymorphism array. A study comparing performance metrics of PRAME IHC and diagnostic GEP has not been previously published. Herein, we evaluated the use of PRAME IHC in 55 melanocytic tumors with challenging histomorphology by comparing the results with diagnostic GEP and final histomorphologic diagnosis. Intertest agreement occurred in 88% of cases. PRAME IHC supported the final diagnosis in 89% of cases with a sensitivity of 79%, specificity of 95%, and positive predictive value of 88.2%. GEP agreed with the final diagnosis in 88% of cases with a sensitivity of 65%, 97% specificity, and positively predicted melanoma in 91.7% of cases. Because the results of this study align with past publications evaluating the performance metrics of PRAME IHC, showing it to be as sensitive as and more cost effective than all other ancillary molecular tests, we propose the use of PRAME IHC as the optimal first-line diagnostic tool for ambiguous melanocytic proliferations.

Pixel Self-Attention Guided Real-Time Instance Segmentation for Group Raised Pigs

Instance segmentation is crucial to modern agriculture and the management of pig farms. In practical farming environments, challenges arise due to the mutual adhesion, occlusion, and dynamic changes in body posture among pigs, making accurate segmentation of multiple target pigs complex. To address these challenges, we conducted experiments using video data captured from varying angles and non-fixed lenses. We selected 45 pigs aged between 20 and 105 days from eight pens as research subjects. Among these, 1917 images were meticulously labeled, with 959 images designated for the training set, 192 for validation, and 766 for testing. To enhance feature utilization and address limitations in the fusion process between bottom-up and top-down feature maps within the feature pyramid network (FPN) module of the YOLACT model, we propose a pixel self-attention (PSA) module, incorporating joint channel and spatial attention. The PSA module seamlessly integrates into multiple stages of the FPN feature extraction within the YOLACT model. We utilized ResNet50 and ResNet101 as backbone networks and compared performance metrics, including AP0.5, AP0.75, AP0.5-0.95, and AR0.5-0.95, between the YOLACT model with the PSA module and YOLACT models equipped with BAM, CBAM, and SCSE attention modules. Experimental results indicated that the PSA attention module outperforms BAM, CBAM, and SCSE, regardless of the selected backbone network. In particular, when employing ResNet101 as the backbone network, integrating the PSA module yields a 2.7% improvement over no attention, 2.3% over BAM, 2.4% over CBAM, and 2.1% over SCSE across the AP0.5-0.95 metric. We visualized prototype masks within YOLACT to elucidate the model’s mechanism. Furthermore, we visualized the PSA attention to confirm its ability to capture valuable pig-related information. Additionally, we validated the transfer performance of our model on a top-down view dataset, affirming the robustness of the YOLACT model with the PSA module.

Association of radiation dose to cardiac substructures with major ischaemic events following breast cancer radiotherapy.

Patients with left-sided breast cancer receive a higher mean heart dose (MHD) after radiotherapy, with subsequent risk of ischaemic heart disease. However, the optimum dosimetric predictor among cardiac substructures has not yet been determined. This study retrospectively reviewed 2158 women with breast cancer receiving adjuvant radiotherapy. The primary endpoint was a major ischaemic event. The dose-volume parameters of each delineated cardiac substructure were calculated. The risk factors for major ischaemic events and the association between MHD and major ischaemic events were analysed by Cox regression. The optimum dose-volume predictors among cardiac substructures were explored in multivariable models by comparing performance metrics of each model. At a median follow-up of 7.9 years (interquartile range 5.6-10.8 years), 89 patients developed major ischaemic events. The cumulative incidence rate of major ischaemic events was significantly higher in left-sided disease (P = 0.044). Overall, MHD increased the risk of major ischaemic events by 6.2% per Gy (hazard ratio 1.062, 95% confidence interval 1.01-1.12; P = 0.012). The model containing the volume of the left ventricle receiving 25 Gy (LV V25) with the cut-point of 4% presented with the best goodness of fit and discrimination performance in left-sided breast cancer. Age, chronic kidney disease, and hyperlipidaemia were also significant risk factors. Risk of major ischaemic events exist in the era of modern radiotherapy. LV V25 ≥ 4% appeared to be the optimum parameter and was superior to MHD in predicting major ischaemic events. This dose constraint could aid in achieving better heart protection in breast cancer radiotherapy, though a further validation study is warranted.

Direct CO2-to-renewable methanol: Outlook, performance and optimization approach

Renewable methanol (MeOH) is targeted as a strategically important feedstock and fuel for the sustainable chemicals and industrial fuels sectors as well as a high-density storage and transport medium for the emerging hydrogen (H2) economy. This study elucidates how to approach the future of the renewable thermal direct‑carbon dioxide (CO2)-to-MeOH process supported by a technoeconomic analysis (TEA) to enable high MeOH yield and 3–14% more affordable productivity. We also carry out thermal performance evaluations of five recognizable catalyst materials that achieve up to 38–112% activity compared to the commercial-benchmark copper‑zinc oxide-alumina (CZA) catalyst at low temperature and 1/10th commercial pressure with the same screening reactor. We outline comparative thermal performance metrics and present an optimization strategy to identify the most promising candidates for further optimization.

The Use of Sustainable Financial Instruments in Relation to the Social Impact Investment: ESG Policies, Capital Markets’ Approach and Investors’ Protection: An Innovative Perspective for a Global Surveillance Authority

Social Impact Investment (SII) refers to the provision of finance to organizations addressing social needs. It is intended to achieve both a social impact as well as garner financial returns. Thus, SII may be understood as a business case for what was previously considered mostly philanthropic activities. In the case of SRI (Sustainable and Responsible Investment), investors decide in which company to invest on the basis of Environmental, Social and Governance (ESG) criteria. They consider a company's performance with respect to its effects on nature, its relationships with employees, clients and society, and the transparency of its governance. Where possible, companies should provide quantitative, comparable and forward-looking performance metrics to facilitate ESG integration. It is recommended that companies should use credible indicators which are included in internationally recognised reporting frameworks like GRI, SASB, IIRC, TCFD, CDP or UNGC. Companies should disclose their methodology and provide explanations to support quantitative indicators (ESG metrics), establishing the link between ESG and financial performance. The investor perspective is in important issue, which should be taken under consideration by European and local supervising authorities, under the provisions of the Market Abuse Directive and the MiFID, especially after the introduction and use of ESG and sustainable financial instruments, in both, Banking and Capital Markets. Therefore, SII can be a very useful tool for achieving EU Social Policy objectives. EU efforts (DG EMPL/FISMA) can be targeted at four specific aims that concern impact measurement: 1) promotion of the standardisation of impact measurement through the further development of the taxonomy regulations to include social objectives or the creation of standardised guidelines or the further development of ESG metrics; 2) capacity building in the area of impact measurement; 3) increasing the availability of the data needed; and 4) the promotion of additional and more specialized sustainable financial instruments linked to the Banking Sector and Capital Markets, which in turn will leverage more the total investment amount derived from the InvestEU tool and offer to all stakeholders more flexibility. Therefore, the global liquidity will be increased, and a strong economic turmoil could be avoided under the surveillance of a theoretically established Global Surveillance Authority. The use of sustainable financial instruments by a P.E.S. (social enterprise) under a public-private partnership (PPPs) is also feasible and socially rewarding.

Combined Surface-Subsurface Stream Restoration Structures Can Optimize Hyporheic Attenuation of Stream Water Contaminants.

There is a design-to-function knowledge gap regarding how engineered stream restoration structures can maximize hyporheic contaminant attenuation. Surface and subsurface structures have each been studied in isolation as techniques to restore hyporheic exchange, but surface-subsurface structures have not been investigated or optimized in an integrated manner. Here, we used a numerical model to systematically evaluate key design variables for combined surface (i.e., weir height and length) and subsurface (i.e., upstream and downstream baffle plate spacing) structures. We also compared performance metrics that place differing emphasis on hyporheic flux versus transit times. We found that surface structures tended to create higher flux, shorter transit time flowpaths, whereas subsurface structures promoted moderate-flux, longer transit time flowpaths. Optimal combined surface-subsurface structures could increase fluxes and transit times simultaneously, thus providing conditions for contaminant attenuation that were many times more effective than surface or subsurface structures alone. All performance metrics were improved by the presence of an upstream plate and the absence of a downstream plate. Increasing the weir length tended to improve all metrics, whereas the optimal weir height varied based on metrics. These findings may improve stream restoration by better aligning specific restoration goals with appropriate performance metrics and hyporheic structure designs.

Safe implementation of surgical innovation: a prospective registry of the Versius Robotic Surgical System

ObjectivesTo describe a new, international, prospective surgical registry developed to accompany the clinical implementation of the Versius Robotic Surgical System by accumulating real-world evidence of its safety and effectiveness.InterventionsThis robotic...

Universal scaling theory of electrochemical immunosensors: An analytical approach to define and compare performance metrics

Electrochemical immunosensors have emerged as a versatile, sensitive, and selective sensor technology of choice for a variety of applications, including detection of proteins, food pathogens, bacteria, viruses, and cancerous molecules. The combination of highly specific biorecognition elements and electrical readout systems facilitates the detection of antigens down to femtomolar concentrations. However, a lack of quantitative theoretical framework has made the design, optimization, and comparison of sensors difficult, without a clear and definitive understanding of the limits of detection, dynamic range, and sensitivity. In this paper, we integrate reaction-diffusion and effective media theories to derive a generalized scaling model for an arbitrary immunosensor that relates the relative change of redox current to the corresponding change in antigen concentration, through scaling exponents related to the geometry of biomolecules diffusion and the measurement resolution. Experimental data from dozens of immunosensors (for a variety of antigens, material systems, and sensor geometry) validate our sensor-agnostic scaling formula. Our results would allow cross-calibration of the emerging and traditional immunosensors reported across the literature and define a physics-based, standardized methodology to compare performance metrics, such as limits of detection, dynamic range, and sensitivity.

Multiple Models in Predicting Acquisitions in the Indian Manufacturing Sector: A Performance Comparison

Aim/Purpose: Acquisitions play a pivotal role in the growth strategy of a firm. Extensive resources and time are dedicated by a firm toward the identification of prospective acquisition candidates. The Indian manufacturing sector is currently experiencing significant growth, organically and inorganically, through acquisitions. The principal aim of this study is to explore models that can predict acquisitions and compare their performance in the Indian manufacturing sector. Background: Mergers and Acquisitions (M&A) have been integral to a firm’s growth strategy. Over the years, academic research has investigated multiple models for predicting acquisitions. In the context of the Indian manufacturing industry, the research is limited to prediction models. This research paper explores three models, namely Logistic Regression, Decision Tree, and Multilayer Perceptron, to predict acquisitions. Methodology: The methodology includes defining the accounting variables to be used in the model which have been selected based on strong theoretical foundations. The Indian manufacturing industry was selected as the focus, specifically, data for firms listed in the Bombay Stock Exchange (BSE) between 2010 and 2022 from the Prowess database. There were multiple techniques, such as data transformation and data scrubbing, that were used to mitigate bias and enhance the data reliability. The dataset was split into 70% training and 30% test data. The performance of the three models was compared using standard metrics. Contribution: The research contributes to the existing body of knowledge in multiple dimensions. First, a prediction model customized to the Indian manufacturing sector has been developed. Second, there are accounting variables identified specific to the Indian manufacturing sector. Third, the paper contributes to prediction modeling in the Indian manufacturing sector where there is limited research. Findings: The study found significant supporting evidence for four of the proposed hypotheses indicating that accounting variables can be used to predict acquisitions. It has been ascertained that statistically significant variables influence acquisition likelihood: Quick Ratio, Equity Turnover, Pretax Margin, and Total Sales. These variables are intrinsically linked with the theories of liquidity, growth-resource mismatch, profitability, and firm size. Furthermore, comparing performance metrics reveals that the Decision Tree model exhibits the highest accuracy rate of 62.3%, specificity rate of 66.4%, and the lowest false positive ratio of 33.6%. In contrast, the Multilayer Perceptron model exhibits the highest precision rate of 61.4% and recall rate of 64.3%. Recommendations for Practitioners: The study findings can help practitioners build custom prediction models for their firms. The model can be developed as a live reference model, which is continually updated based on a firm’s results. In addition, there is an opportunity for industry practitioners to establish a benchmark score that provides a reference for acquisitions. Recommendation for Researchers: Researchers can expand the scope of research by including additional classification modeling techniques. The data quality can be enhanced by cross-validation with other databases. Textual commentary about the target firms, including management and analyst quotes, provides additional insight that can enhance the predictive power of the models. Impact on Society: The research provides insights into leveraging emerging technologies to predict acquisitions. The theoretical basis and modeling attributes provide a foundation that can be further expanded to suit specific industries and firms. Future Research: There are opportunities to expand the scope of research in various dimensions by comparing acquisition prediction models across industries and cross-border and domestic acquisitions. Additionally, it is plausible to explore further research by incorporating non-financial data, such as management commentary, to augment the acquisition prediction model.

Evaluating models for classifying movement of whale-watching vessels

A continued rise in global ocean vessel activity has led to growing concerns for the health of whales around the world. Of particular interest is the increase in recreation vessels, including those related to whale-watching activities. However, there is an absence of established procedures to identify vessels engaged in whale-watching, thus limiting the ability to quantify whale-watching impacts on whales. This study evaluates three computational classification models and their ability to utilize Automatic Identification System (AIS) data to describe wildlife-viewing vessel behaviour. These models include a density-based spatial clustering application with noise (DBSCAN), a hidden Markov model (HMM), and logistic regression (LR), all of which have been previously used to classify vessel behaviour in industries, such as fishing, shipping, and marine security. The results of each model's classification were validated against observed whale sighting data using statistical performance and accuracy metrics. The findings suggest that all three classification models sufficiently detect wildlife-viewing behaviour, but the HMM and LR had preferable performance metrics compared to DBSCAN. Further, although LR provides an informative glance at which AIS variables are most important to detecting wildlife-viewing events, the HMM has comparable performance metrics and requires less data processing. Therefore, this study recommends the use of HMM due to its computational efficiency and because it provides an accurate classification of wildlife-viewing behaviour for whale-watching vessels. The results of this study can be used to support policy decisions, monitor regulation compliance, and inform marine conservation initiatives.

Efficient Model for Coronary Artery Disease Diagnosis: A Comparative Study of Several Machine Learning Algorithms.

Background In today's industrialized world, coronary artery disease (CAD) is one of the leading causes of death, and early detection and timely intervention can prevent many of its complications and eliminate or reduce the resulting mortality. Machine learning (ML) methods as one of the cutting-edge technologies can be used as a suitable solution in diagnosing this disease. Methods In this study, different ML algorithms' performances were compared for their effectiveness in developing a model for early CAD diagnosis based on clinical examination features. This applied descriptive study was conducted on 303 records and overall 26 features, of which 26 were selected as the target features with the advice of several clinical experts. In order to provide a diagnostic model for CAD, we ran most of the most critical classification algorithms, including Multilayer Perceptron (MLP), Support Vector Machine (SVM), Logistic Regression (LR), J48, Random Forest (RF), K-Nearest Neighborhood (KNN), and Naive Bayes (NB). Seven different classification algorithms with 26 predictive features were tested to cover all feature space and reduce model error, and the most efficient algorithms were identified by comparison of the results. Results Based on the compared performance metrics, SVM (AUC = 0.88, F-measure = 0.88, ROC = 0.85), and RF (AUC = 0.87, F-measure = 0.87, ROC = 0.91) were the most effective ML algorithms. Among the algorithms, the KNN algorithm had the lowest efficiency (AUC = 0.81, F-measure = 0.81, ROC = 0.77). In the diagnosis of coronary artery disease, machine learning algorithms have played an important role. Proposed ML models can provide practical, cost-effective, and valuable support to doctors in making decisions according to a good prediction. Discussion. It can become the basis for developing clinical decision support systems. SVM and RF algorithms had the highest efficiency and could diagnose CAD based on patient examination data. It is suggested that further studies be performed using these algorithms to diagnose coronary artery disease to obtain more accurate results.

Predictive Models for Length of Stay and Discharge Disposition in Elective Spine Surgery: Development, Validation, and Comparison to the ACS NSQIP Risk Calculator.

A retrospective study at a single academic institution. The purpose of this study is to utilize machine learning to predict hospital length of stay (LOS) and discharge disposition following adult elective spine surgery, and to compare performance metrics of machine learning models to the American College of Surgeon's National Surgical Quality Improvement Program's (ACS NSQIP) prediction calculator. A total of 3678 adult patients undergoing elective spine surgery between 2014 and 2019, acquired from the electronic health record. Patients were divided into three stratified cohorts: cervical degenerative, lumbar degenerative, and adult spinal deformity groups. Predictive variables included demographics, body mass index, surgical region, surgical invasiveness, surgical approach, and comorbidities. Regression, classification trees, and least absolute shrinkage and selection operator (LASSO) were used to build predictive models. Validation of the models was conducted on 16% of patients (N=587), using area under the receiver operator curve (AUROC), sensitivity, specificity, and correlation. Patient data were manually entered into the ACS NSQIP online risk calculator to compare performance. Outcome variables were discharge disposition (home vs. rehabilitation) and LOS (days). Of 3678 patients analyzed, 51.4% were male (n=1890) and 48.6% were female (n=1788). The average LOS was 3.66 days. In all, 78% were discharged home and 22% discharged to rehabilitation. Compared with NSQIP (Pearson R2 =0.16), the predictions of poisson regression ( R2 =0.29) and LASSO ( R2 =0.29) models were significantly more correlated with observed LOS ( P =0.025 and 0.004, respectively). Of the models generated to predict discharge location, logistic regression yielded an AUROC of 0.79, which was statistically equivalent to the AUROC of 0.75 for NSQIP ( P =0.135). The predictive models developed in this study can enable accurate preoperative estimation of LOS and risk of rehabilitation discharge for adult patients undergoing elective spine surgery. The demonstrated models exhibited better performance than NSQIP for prediction of LOS and equivalent performance to NSQIP for prediction of discharge location.

Predicting Time to Death After Withdrawal of Life-Sustaining Treatment in Children.

OBJECTIVES:Accurately predicting time to death after withdrawal of life-sustaining treatment is valuable for family counseling and for identifying candidates for organ donation after cardiac death. This topic has been well studied in adults, but literature is scant in pediatrics. The purpose of this report is to assess the performance and clinical utility of the available tools for predicting time to death after treatment withdrawal in children.DATA SOURCES:Terms related to predicting time to death after treatment withdrawal were searched in PubMed and Embase from 1993 to November 2021.STUDY SELECTION:Studies endeavoring to predict time to death or describe factors related to time to death were included. Articles focusing on perceptions or practices of treatment withdrawal were excluded.DATA EXTRACTION:Titles, abstracts, and full text of articles were screened to determine eligibility. Data extraction was performed manually. Two-by-two tables were reconstructed with available data from each article to compare performance metrics head to head.DATA SYNTHESIS:Three hundred eighteen citations were identified from the initial search, resulting in 22 studies that were retained for full-text review. Among the pediatric studies, predictive models were developed using multiple logistic regression, Cox proportional hazards, and an advanced machine learning algorithm. In each of the original model derivation studies, the models demonstrated a classification accuracy ranging from 75% to 91% and positive predictive value ranging from 0.76 to 0.93.CONCLUSIONS:There are few tools to predict time to death after withdrawal of life-sustaining treatment in children. They are limited by small numbers and incomplete validation. Future work includes utilization of advanced machine learning models.

Career Research Performance Among Radiology Early Career Grant Recipients Compared With National Institutes of Health K Award Recipients.

To compare academic and demographic metrics among recipients of three major early career radiology, interventional radiology, and radiation oncology grants to National Institutes of Health (NIH) K awardees at the time the grants were awarded and then over the course of their careers. Radiologists who received the RSNA Research Scholar Grant, General Electric Radiology Research Academic Fellowship (GERRAF), American Roentgen Ray Society (ARRS) Scholar Award, or NIH Career Development (K) Award before January 1, 2015, were included. Research metrics at the time of grant award (eg, publications) and subsequent scholarly productivity (eg, academic rank, h-index, NIH funding) were recorded until April 2020. Wilcoxon ranked-sum, χ2, logistic regression, and standard least-square regression were used for data analysis. There were 279 recipients: 48 K Award, 115 RSNA Research Scholar Grant, 36 ARRS, and 80 GERRAF. At the time of grant awarding, GERRAF recipients were less likely to have an MD-PhD degree (odds ratio [OR]: 0.36; P= .002) and were more likely to be women (OR: 1.55; P= .042) than K Award recipients. Similarly, recipients of the ARRS (OR: 2.87; P= .010) and GERRAF (OR: 3.19; P= .002) were more likely to have a master's degree. Academic rank, leadership positions, and R01 funding were significant predictors of h-index and total publications over time. Academic rank and the GERRAF were significant predictors of subsequent NIH funding duration but there were no significant predictors of NIH funding amount. Early career radiology awards, specifically the GERRAF, provide support for female and non-PhD investigators and result in comparable academic performance metrics to NIH K Award recipients.