ROC AUC Research Articles

Comparative analysis of groundwater vulnerability in the Oued Agrioun watershed, northeast Algeria: DRASTIC, GOD, and SINTACS approaches with ROC AUC validation

This study focuses on the Oued Agrioun watershed, a semi-arid region in northeastern Algeria, with the aim of mapping the vulnerability to pollution in order to improve groundwater management. To achieve this, a comprehensive database was compiled from a variety of sources, including geological, hydrogeological, geomorphological, and climatic data, as well as satellite imagery. Three different methods are applied using Geographic Information Systems (GIS): DRASTIC, SINTACS, and GOD. The GOD and SINTACS methods resulted in four main classes of vulnerability: low, medium, high, and very high. In contrast, the DRASTIC model identified three classes: low (2.04%), medium (45.04%) and high (52.93%). ROC curve validation using nitrate concentrations indicates that the SINTACS and GOD methods provide better predictive accuracy than DRASTIC, suggesting that a combination of these approaches provides a more robust assessment of the vulnerability of subterranean waters. This study provides valuable data to guide future policy decisions aimed at protecting groundwater resources in the Oued Agrioun watershed.This study focuses on the Oued Agrioun watershed, a semi-arid region in northeastern Algeria, with the aim of mapping vulnerability to pollution in order to improve groundwater management. To achieve this, a comprehensive database was compiled from a variety of sources, including geological, hydrogeological, geomorphological, and climatic data, as well as satellite imagery. Three different methods are applied using Geographic Information Systems (GIS): DRASTIC, SINTACS, and GOD. The GOD and SINTACS methods resulted in four main classes of vulnerability: low, medium, high, and very high. In contrast, the DRASTIC model identified three classes: low (2.04%), medium (45.04%), and high (52.93%). Validation by ROC curve using nitrate concentrations indicates that the SINTACS and GOD methods provide better predictive accuracy than DRASTIC, suggesting that a combination of these approaches provides a more robust assessment of the vulnerability of groundwater resources. This study provides valuable data to guide future policy decisions aimed at protecting groundwater resources in the Oued Agrioun watershed.

Intersection of Performance, Interpretability, and Fairness in Neural Prototype Tree for Chest X-Ray Pathology Detection: Algorithm Development and Validation Study.

While deep learning classifiers have shown remarkable results in detecting chest X-ray (CXR) pathologies, their adoption in clinical settings is often hampered by the lack of transparency. To bridge this gap, this study introduces the neural prototype tree (NPT), an interpretable image classifier that combines the diagnostic capability of deep learning models and the interpretability of the decision tree for CXR pathology detection. This study aimed to investigate the utility of the NPT classifier in 3 dimensions, including performance, interpretability, and fairness, and subsequently examined the complex interaction between these dimensions. We highlight both local and global explanations of the NPT classifier and discuss its potential utility in clinical settings. This study used CXRs from the publicly available Chest X-ray 14, CheXpert, and MIMIC-CXR datasets. We trained 6 separate classifiers for each CXR pathology in all datasets, 1 baseline residual neural network (ResNet)-152, and 5 NPT classifiers with varying levels of interpretability. Performance, interpretability, and fairness were measured using the area under the receiver operating characteristic curve (ROC AUC), interpretation complexity (IC), and mean true positive rate (TPR) disparity, respectively. Linear regression analyses were performed to investigate the relationship between IC and ROC AUC, as well as between IC and mean TPR disparity. The performance of the NPT classifier improved as the IC level increased, surpassing that of ResNet-152 at IC level 15 for the Chest X-ray 14 dataset and IC level 31 for the CheXpert and MIMIC-CXR datasets. The NPT classifier at IC level 1 exhibited the highest degree of unfairness, as indicated by the mean TPR disparity. The magnitude of unfairness, as measured by the mean TPR disparity, was more pronounced in groups differentiated by age (chest X-ray 14 0.112, SD 0.015; CheXpert 0.097, SD 0.010; MIMIC 0.093, SD 0.017) compared to sex (chest X-ray 14 0.054 SD 0.012; CheXpert 0.062, SD 0.008; MIMIC 0.066, SD 0.013). A significant positive relationship between interpretability (ie, IC level) and performance (ie, ROC AUC) was observed across all CXR pathologies (P<.001). Furthermore, linear regression analysis revealed a significant negative relationship between interpretability and fairness (ie, mean TPR disparity) across age and sex subgroups (P<.001). By illuminating the intricate relationship between performance, interpretability, and fairness of the NPT classifier, this research offers insightful perspectives that could guide future developments in effective, interpretable, and equitable deep learning classifiers for CXR pathology detection.

Diagnostic accuracy of artificial intelligence for the screening of prostate cancer in biparametric magnetic resonance imaging: a systematic review

BACKGROUND: Based on the latest published data, 40,137 new cases of prostate cancer were reported in Russia in 2021, ranking second after lung cancer in men. Thus, prostate cancer is one of the most common malignant neoplasms in men. Accurate and timely detection of prostate cancer is important under the current conditions. AIM: This systematic review aimed to assess the quality of prediction models designed to detect prostate cancer during initial presentation. MATERIALS AND METHODS: A systematic search was performed in eLibrary.ru, PubMed, Google Scholar, Web of Science, and ResearchGate for relevant publications indexed from January 2019 to September 2023 in accordance with the PRISMA protocol. Two authors independently assessed the relevant studies for potential inclusion or exclusion. RESULTS: This systematic review meta-analysis included 21 studies. In total, data from 3,630 patients were analyzed, of which 47% had prostate cancer and 53% had benign prostate neoplasms. The mean age of the patients was 67.1 (36–90) years. In addition, 81% of the studies were based on T2-weighted imaging, 57% on diffusion-weighted imaging, and 76% on apparent diffusion coefficient. Moreover, 43% and 33% of the studies were dedicated to transition zone and prostate peripheral zone neoplasms, respectively, and 52% of the authors examined the whole prostate gland, without dividing it into zones. The most common machine-learning algorithms applied by the investigators were as follows: multiple logistic regression (76%), support vector machine (38%), and random forest (24%). Based on the meta-analysis performed for the receiver operating characteristic-area under the curve (ROC–AUC) assessment with random-effect approach in 73 prediction models described in the publications, the final ROC–AUC was 0.793 [95% CI 0.768–0.818], I2 = 86.71%, p 0.001. The most accurate prediction models were based on the T2-weighted imaging + apparent diffusion coefficients imaging protocol: 0.860 [95% CI 0.813–0.907], and models created according to the “white box” principle (0.834 [95% CI 0.806–0.861]) were more accurate than the “black box” ones (0.733 [95% CI 0.695–0.771]). The models using radiomics and clinical features were slightly more accurate than those using the radiomics parameters alone (0.869 [95% CI 0.844–0.895] vs. 0.779 [95% CI 0.751–0.807]). Model accuracy was nearly identical across transitional and/or peripheral zone studies. CONCLUSIONS: Artificial intelligence demonstrated promising results. However, the clinical applicability may require more intensive expert inspection in healthcare institutions and evaluation of efficacy in prospective studies.

Elevating fraud detection: machine learning models with computational intelligence optimization

&lt;span lang="EN-US"&gt;The amount of crimes committed online has undoubtedly increased as more people use the internet for e-commerce and other financial transactions. Machine learning algorithms have been created to detect payment fraud in online purchasing in order to address the issue. This study performs a thorough comparative examination of different metaheuristic optimizations as hyperparameter tuning methods; these are Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). They are used to optimize the ROC AUC of the three machine learning algorithms, namely X-Gradient Boost, Random Forest Classifier, and Light Gradient Boost Machine. Since the study's data are unbalanced, the determined metrics were ROC AUC. PSO offers consistent conditions for finding the best solution, according to our experiment. Without the inclusion of population annihilation strategies, PSO can achieve the greatest results in various situations which are different from GA, a consistent condition for finding the best solution, according to our experiment. Without the inclusion of population annihilation strategies, PSO can achieve the greatest results in various situations. The findings indicate that Random Forest Classifier provided the highest ROC AUC value both before and after the hyperparameter tuning process, with a score of 88.69% attained while utilizing PSO.&lt;/span&gt;

The detection of GRN mutation carriers by progranulin blood protein levels from finger-stick collection.

Heterozygous mutations in the progranulin gene (GRN) leading to decreased progranulin levels are one of the most frequent causes of inherited frontotemporal dementia (FTD). We evaluated progranulin levels in dried blood spots from capillary finger-stick collection (DBScapillary). Paired venous Ethylenediaminetetraacetic acid (EDTA) plasma and DBScapillary samples were collected from each participant with or without pathogenic GRN mutations. DBScapillary progranulin levels in GRN mutation carriers (mean [SD] age, 55 [13] years; n=16) were reduced compared to non-mutation carriers (64 [11] years; n=44) (2.38ng/mL [1.0] vs 4.37 [0.68] ng/mL; U=42; p<0.0001, ROC AUC=0.94 [95% CI: 0.83 to 1.00]) and highly associated with venous plasma levels (R=0.819; p<0.001). Progranulin levels can be accurately determined from finger-stick blood samples. This can enable regular and remote monitoring of this protein in FTD therapeutic trials and potentially serve as a first-level screening test for GRN mutations. Progranulin levels measured using capillary dried blood spots were significantly reduced in GRN mutation carriers compared to non-mutation carriers. Progranulin levels measured using capillary dried blood spots strongly correlated with levels from venous EDTA plasma. DBScapillary progranulin levels were able to identify GRN mutation carriers with high accuracy. DBScapillary might allow repeated measurements of progranulin levels in a remote and unsupervised setting, circumventing the restrictions of traditional venous blood collection. DBScapillary might be used to assess the biological efficacy of disease-modifying therapies in clinical trials aiming to increase baseline progranulin levels or as a first-level screening for GRN mutations in primary settings.

Targeted Proteomic Biomarker Profiling Using NULISA in a cohort enriched with risk for Alzheimer's Disease and Related Dementias.

Targeted proteomic assays may be useful for diagnosing and staging Alzheimer's disease and related dementias (ADRD). We evaluated the performance of a 120-marker central nervous system (CNS) NUcleic acid-Linked Immuno-Sandwich Assay (NULISA) panel in samples spanning the AD spectrum. Cross-sectional plasma samples (n=252) were analyzed using Alamar's NULISAseq CNS panel. ROC analyses demonstrated NULISAseq-pTau217 accuracy in detecting amyloid (A) and tau (T) PET positivity. Differentially expressed proteins were identified using volcano plots. NULISAseq-pTau217 accurately classified A/T PET status with ROC AUCs of 0.92/0.86. pTau217 was upregulated in A+, T+, and impaired groups with log2-fold changes of 1.21, 0.57 and 4.63, respectively, compared to A-. Interestingly, pTDP43-409 was also upregulated in the impaired group and correlated with declining hippocampal volume and cognitive trajectories. This study shows the potential of a targeted proteomics panel for characterizing brain changes pertinent to ADRD. The promising pTDP43-409 findings require further replication.

Modeling human decomposition: A Bayesian approach

Environmental and individualistic variables affect the rate of human decomposition in complex ways. These effects complicate the estimation of the postmortem interval (PMI) based on observed decomposition characteristics. In this work, we develop a generative probabilistic model for decomposing human remains based on PMI and a wide range of environmental and individualistic variables. This model explicitly represents the effect of each variable, including PMI, on the appearance of each decomposition characteristic, allowing for direct interpretation of model effects and enabling the use of the model for PMI inference and optimal experimental design. In addition, the probabilistic nature of the model allows for the integration of expert knowledge in the form of prior distributions. We fit this model to a diverse set of 2529 cases from the GeoFOR dataset. We demonstrate that the model accurately predicts 24 decomposition characteristics with an ROC AUC score of 0.85. Using Bayesian inference techniques, we invert the decomposition model to predict PMI as a function of the observed decomposition characteristics and environmental and individualistic variables, producing an R-squared measure of 71 %. Finally, we demonstrate how to use the fitted model to design future experiments that maximize the expected amount of new information about the mechanisms of decomposition using the Expected Information Gain formalism.

ParaAntiProt provides paratope prediction using antibody and protein language models

Efficiently predicting the paratope holds immense potential for enhancing antibody design, treating cancers and other serious diseases, and advancing personalized medicine. Although traditional methods are highly accurate, they are often time-consuming, labor-intensive, and reliant on 3D structures, restricting their broader use. On the other hand, machine learning-based methods, besides relying on structural data, entail descriptor computation, consideration of diverse physicochemical properties, and feature engineering. Here, we develop a deep learning-assisted prediction method for paratope identification, relying solely on amino acid sequences and being antigen-agnostic. Built on the ProtTrans architecture, and utilizing pre-trained protein and antibody language models, we extract efficient embeddings for predicting paratope. By incorporating positional encoding for Complementarity Determining Regions, our model gains a deeper structural understanding, achieving remarkable performance with a 0.904 ROC AUC, 0.701 F1-score, and 0.585 MCC on benchmark datasets. In addition to yielding accurate antibody paratope predictions, our method exhibits strong performance in predicting nanobody paratope, achieving a ROC AUC of 0.912 and a PR AUC of 0.665 on the nanobody dataset. Notably, our approach outperforms structure-based prediction methods, boasting a PR AUC of 0.731. Various conducted ablation studies, which elaborate on the impact of each part of the model on the prediction task, show that the improvement in prediction performance by applying CDR positional encoding together with CNNs depends on the specific protein and antibody language models used. These results highlight the potential of our method to advance disease understanding and aid in the discovery of new diagnostics and antibody therapies.

Ensemble Voting Method for Phonocardiogram Heart Signal Classification Using FFT Features

Heart disease is still one of the leading causes of death worldwide, hence the need for effective diagnostic tools. Phonocardiogram (PCG) signals have been explored as a complementary approach to electrocardiogram (ECG) to detect cardiac abnormalities. This research investigates the classification of PCG signals using Fast Fourier Transform (FFT) features and deep learning models, including Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Temporal Convolutional Network (TCN). Hyperparameter tuning, particularly learning rate adjustment, is applied to optimize the performance of the models. The results show that the GRU and TCN models outperform the LSTM, achieving up to 92% accuracy at a learning rate of 0.0001. Ensemble learning with soft voting was also applied to combine the strengths of each model. Although the ensemble model showed strong performance with 92% accuracy and ROC AUC of 0.9636, it did not provide significant improvement over the base model. This finding highlights the importance of hyperparameter tuning in model optimization, with GRU and TCN showing slightly better performance in the time series classification task. This study concludes that ensemble learning offers stability but does not significantly improve classification accuracy beyond a well-tuned base model.

Ion channel classification through machine learning and protein language model embeddings.

Ion channels are critical membrane proteins that regulate ion flux across cellular membranes, influencing numerous biological functions. The resource-intensive nature of traditional wet lab experiments for ion channel identification has led to an increasing emphasis on computational techniques. This study extends our previous work on protein language models for ion channel prediction, significantly advancing the methodology and performance. We employ a comprehensive array of machine learning algorithms, including k-Nearest Neighbors, Random Forest, Support Vector Machines, and Feed-Forward Neural Networks, alongside a novel Convolutional Neural Network (CNN) approach. These methods leverage fine-tuned embeddings from ProtBERT, ProtBERT-BFD, and MembraneBERT to differentiate ion channels from non-ion channels. Our empirical findings demonstrate that TooT-BERT-CNN-C, which combines features from ProtBERT-BFD and a CNN, substantially surpasses existing benchmarks. On our original dataset, it achieves a Matthews Correlation Coefficient (MCC) of 0.8584 and an accuracy of 98.35 %. More impressively, on a newly curated, larger dataset (DS-Cv2), it attains an MCC of 0.9492 and an ROC AUC of 0.9968 on the independent test set. These results not only highlight the power of integrating protein language models with deep learning for ion channel classification but also underscore the importance of using up-to-date, comprehensive datasets in bioinformatics tasks. Our approach represents a significant advancement in computational methods for ion channel identification, with potential implications for accelerating research in ion channel biology and aiding drug discovery efforts.

The predictive value of CK-MB on the curative effect of urokinase combined with LMWH calcium in AMI.

This study investigates the predictive value of CK-MB in treating acute myocardial infarction (AMI) with urokinase combined with low molecular weight heparin (LMWH) calcium. AMI patients treated at our hospital from April 2020 to August 2022 were included in this prospective, multi-center cohort study. Patients were randomly assigned to a urokinase group or a combined treatment group (urokinase plus LMWH). CK-MB levels were measured using ELISA, and adverse reactions were recorded. The accuracy of CK-MB as a predictor was evaluated via ROC curve analysis and AUC. Before treatment, CK-MB levels were similar in both groups (p > 0.05). Post-treatment, levels were significantly lower in the combined group (p < 0.05). The combined group had a higher proportion of mild CK-MB elevation and lower moderate/severe elevations (p < 0.05). Adverse reaction rates were similar (p > 0.05). The combined treatment group showed higher effectiveness (p < 0.05), with better sensitivity (94%) and specificity (82%) than the urokinase group (sensitivity 89%, specificity 78%). CK-MB is a valuable predictor for evaluating the efficacy of urokinase plus LMWH in AMI treatment, offering advantages in reducing CK-MB levels and improving outcomes.

Development of a metabolome-based respiratory infection prognostic during COVID-19 arrival.

In a new respiratory virus pandemic, optimizing allocation of scarce medical resources becomes an urgent challenge. Infection prognosis takes on particular importance when allocating scarce antiviral antibodies and drugs, which are most effective when administered before the onset of severe disease. During arrival of the COVID-19 pandemic to the United States in 2020, we conducted a prognostic biomarker discovery and validation effort based upon metabolomic profiling with a liquid-chromatography-mass spectrometer (LC-MS) type used clinically for rapid toxicology. We obtained urine specimens from 163 patients presenting for evaluation. We obtained LC-MS profiles in the initial cohort and used machine learning methods to define a simplified urine metabolomic signature associated with respiratory failure or death by 90 days. This signature was composed of three metabotypes linked to intestinal microbiome metabolism and anticonvulsant use, with a receiver-operator characteristic area under the curve (ROC AUC) of 89.4%. Blinded application of this signature to the subsequent validation cohort yielded a ROC AUC of 81.2%. A model trained on the two baseline metabotypes present before intubation exhibited similar performance in the validation cohort. This study demonstrates the plausibility and promise of rapid metabolome-based prognostic discovery and validation in the opening wave of a pandemic. The approach used here could be used to inform therapeutic and resource allocation decisions early in a future epidemic.IMPORTANCEIn a new respiratory virus pandemic, the ability to identify patients at greatest risk for severe disease is essential to direct scarce medical resources to those most likely to benefit from them. Tools to predict disease severity are best developed early in a pandemic, but laboratory-based resources to develop these may be limited by available technology and by infection precautions. Here, we show that an accessible metabolic profiling approach could identify a prognostic signature of severe disease in the initial wave of COVID-19, when patients presenting for care often exceeded the available doses of convalescent plasma and remdesivir. In a future pandemic, this approach, alongside efforts to identify clinical disease severity predictors, could improve patient outcomes and facilitate therapeutic trials by identifying individuals at high risk for severe disease.

LncRNAs SOX2-OT and NEAT1 act as a potential biomarker for esophageal squamous cell carcinoma

Despite strides in diagnostic and therapeutic approaches for ESCC, patient survival rates remain relatively low. Recent studies highlight the pivotal role of long non-coding RNAs (lncRNAs) in regulating diverse cellular activities in humans. Dysregulated lncRNAs have emerged as potential diagnostic indicators across various cancers, including ESCC. However, further research is necessary to effectively leverage ESCC-associated lncRNAs in clinical settings. Understanding their clinical significance for ESCC diagnosis and their mechanisms can pave the way for more effective therapeutic strategies. Our qRT-PCR analysis revealed significant downregulation of SOX2-OT (~ 2.02-fold) and NEAT1 (~ 1.53-fold) in ESCC blood samples. These lncRNAs show potential as biomarkers for distinguishing ESCC patients from healthy individuals, with ROC curves and AUC values of 0.736 for SOX2-OT and 0.621 for NEAT1. Further analysis examined the correlation between SOX2-OT and NEAT1 expression and various clinicopathological factors, including age, gender, smoking, alcohol use, hot beverage intake, tumor grade, and TNM stages. In-silico studies highlighted their roles in miRNA sponging via mTOR and MAPK pathways, while co-expression network analysis identified associated genes. This research paves the way for future studies on ESCC prognosis using SOX2-OT and NEAT1 as predictive markers. By thoroughly investigating the functions of these lncRNAs, we aim to deepen our understanding of their potential as diagnostic markers and their role in facilitating effective therapeutic interventions for esophageal squamous cell carcinoma (ESCC) within clinical contexts.

Gut metatranscriptomics based de novo assembly reveals microbial signatures predicting immunotherapy outcomes in non-small cell lung cancer.

Advanced-stage non-small cell lung cancer (NSCLC) poses treatment challenges, with immune checkpoint inhibitors (ICIs) as the main therapy. Emerging evidence suggests the gut microbiome significantly influences ICI efficacy. This study explores the link between the gut microbiome and ICI outcomes in NSCLC patients, using metatranscriptomic (MTR) signatures. We utilized a de novo assembly-based MTR analysis on fecal samples from 29 NSCLC patients undergoing ICI therapy, segmented according to progression-free survival (PFS) into long (> 6months) and short (≤ 6months) PFS groups. Through RNA sequencing, we employed the Trinity pipeline for assembly, MMSeqs2 for taxonomic classification, DESeq2 for differential expression (DE) analysis. We constructed Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) machine learning (ML) algorithms and comprehensive microbial profiles. We detected no significant differences concerning alpha-diversity, but we revealed a biologically relevant separation between the two patient groups in beta-diversity. Actinomycetota was significantly overrepresented in patients with short PFS (vs long PFS, 36.7% vs. 5.4%, p < 0.001), as was Euryarchaeota (1.3% vs. 0.002%, p = 0.009), while Bacillota showed higher prevalence in the long PFS group (66.2% vs. 42.3%, p = 0.007), when comparing the abundance of corresponding RNA reads. Among the 120 significant DEGs identified, cluster analysis clearly separated a large set of genes more active in patients with short PFS and a smaller set of genes more active in long PFS patients. Protein Domain Families (PFAMs) were analyzed to identify pathways enriched in patient groups. Pathways related to DNA synthesis and Translesion were more enriched in short PFS patients, while metabolism-related pathways were more enriched in long PFS patients. E. coli-derived PFAMs dominated in patients with long PFS. RF, SVM and XGBoost ML models all confirmed the predictive power of our selected RNA-based microbial signature, with ROC AUCs all greater than 0.84. Multivariate Cox regression tested with clinical confounders PD-L1 expression and chemotherapy history underscored the influence of n = 6 key RNA biomarkers on PFS. According toML models specific gut microbiome MTR signatures' associate with ICI treated NSCLC outcomes. Specific gene clusters and taxa MTR gene expression might differentiate long vs shortPFS.

Circulating blood circular RNA in Parkinson's Disease; from involvement in pathology to diagnostic tools in at-risk individuals.

To identify circRNAs associated with Parkinson's disease (PD) we leveraged two of the largest publicly available studies with longitudinal clinical and blood transcriptomic data. We performed a cross-sectional study utilizing the last visit of each participant (N = 1848), and a longitudinal analysis that included 1166 participants with at least two time points. We identified 192 differentially expressed circRNAs, with effects that were sustained during disease, in mutation carriers, and diverse ancestry. The 192 circRNAs were leveraged to distinguish between PD and healthy participants with a ROC AUC of 0.797. Further, 71 circRNAs were sufficient to distinguish between genetic PD (AUC71 = 0.954) and, at-risk participants (AUC71 = 0.929) and healthy controls, supporting that circRNAs have the potential to aid the diagnosis of PD. Finally, we identified five circRNAs highly correlated with symptom severity. Overall, we demonstrated that circRNAs play an important role in PD and can be clinically relevant to improve diagnostic and monitoring.

Optimizing Procurement Efficiency by Implementing K-Means and Random Forest in Kopegtel Samarinda’s Warehouse System

Procurement is a company’s activity to purchase goods or equipment needed in operations. In the management process, a procurement management system is often used to facilitate this management, such as at CV Indocitra Multi Artha, which uses the “Sistem Warehouse Kopegtel Samarinda.” The system provides significant assistance to the company, but large requests can be overwhelming to be handled by the manager and can cause an overload information problem. Research was conducted to deal with these problems by implementing a data mining algorithm as a procurement recommendation system. K-means and Random Forest algorithms were chosen as methods for the research. The algorithm is processed within two critical steps, first by K-Means to get cluster data and then by predicting it with Random Forest to get a recommendation for whether the object should be bought or not. Hyperparameter tuning was performed to optimize the model’s performance, yielding an F1-Score of 86.95%, representing the balance between precision and recall, and an ROC AUC value of 82.34%. These substantial metric outcomes indicate that the model can provide practical recommendations

The NACA score predicts mortality in polytrauma patients before hospital admission: a registry-based study

BackgroundThe early assessment of the severity of polytrauma patients is key for their optimal management. The aim of this study was to investigate the discriminative performance of the NACA score in a large dataset by stratifying the severity of polytraumatized patients in correlation to injury severity score (ISS), Glasgow Coma Scale (GCS), and mortality.MethodsThis study on the Swiss Trauma Registry investigated 2239 polytraumatized patient (54.3 ± 22.8 years) enrolled from 2015 to 2023: 0.5% were NACA 3, 76.7% NACA 4, 21.4% NACA 5, and 1.4% NACA 6. The NACA predictive value of patients’ mortality was investigated, as well as the correlation of ISS and GCS scores, and other factors influencing patients’ survival at discharge and after 28 days.ResultsIn NACA 4 and 5 the survival rate during hospitalization was 97.7% and 82.5%, respectively, and 28-day mortality 3.5% and 23.5%, respectively (p < 0.0005). NACA correlated with GCS in the prehospital phase and in the emergency room (p < 0.0005), as well as with ISS (p < 0.0005). NACA 4 and 5 presented different injury patterns (fall < 3 m vs vehicle accident) with NACA 5 requiring more CPR and intubation (p < 0.001, p < 0.0005). The ROC AUC analysis showed the prehospital NACA and GCS values as the strongest variables predicting patients’ survival.ConclusionsThis study provides valuable evidence supporting the effectiveness of the NACA score in assessing the severity of polytrauma patients in both the pre-ER and ER condition. Considering the statistical significant correlation with the GCS and with the ISS, NACA is a valid score for assessing polytrauma patients.

Abstract A075: Circulating T-cell receptor repertoire analysis improves cancer early detection

Abstract Introduction: Cancer screening by liquid biopsy promises to detect cancer early when it can be cured. However, current ctDNA-based approaches detect only a minority of early-stage cancers. For liquid biopsy to realize its full potential for cancer early detection, sensitivity for earliest stage disease must be improved. Methods: We set out to improve sensitivity of liquid biopsy by exploiting tumor recognition by T cells through sequencing of the circulating T-cell receptor repertoire. Using gDNA extracted from blood buffy coats, we studied a cohort of 471 lung cancer patients (85% stage I) and 600 subjects without cancer enriched for older individuals with a history of smoking. We performed TCR beta chain sequencing to yield a median of 101,899 TCR clonotypes per sample (median 81,602 cancer / 112,966 controls) and built a TCR sequence similarity graph to cluster clonotypes into TCR repertoire functional units (RFUs). The TCR frequencies of RFUs were tested for association with cancer status and significantly associated RFUs were combined using an SVM model into a cancer score. Cancer RFU levels were also correlated to imputed subject HLA types for 32 HLA Class I and 38 Class II alleles. The model was evaluated by 10-fold cross-validation and compared to a ctDNA panel of 237 mutation hotspots in 154 lung cancer driver genes and to 17 lung cancer related protein biomarkers in 86 subjects. Results: We identified 372 cancer-associated TCR RFUs at FDR≤0.1, including 198 enriched in cancer samples (fold-change range 1.03 to 3.13) and 174 enriched in controls (fold- change 0.96 to 0.30). Levels of 265/372 (71%) RFUs were correlated to presence of an HLA allele at FDR ≤ 0.1. Of the total 2,770 significant HLA-RFU associations, 479 (17%) involved an HLA Class I allele, while 2,291 (83%) involved an HLA Class II allele, highlighting a potential role for CD4 helper or regulatory T cell antigen recognition in the cancer RFUs found. The RFU cancer score achieved a cross-validated test ROC AUC of 0.71 for cancer status prediction (0.71 Stage I / 0.70 Stage II-IV) with 49% of Stage I cancers detected at a specificity of 80%. Importantly, the cancer score was not dependent on sample source site, technical or biological variation in TCR repertoire depth, or lung cancer histology subtype. The TCR RFU score also distinguished lung cancer patients from control subjects with other conditions such as benign lung nodules and COPD. We observed an up to ∼20%-point gain in sensitivity for stage I cancer when TCR RFUs were added to ctDNA and proteins in a potential multi-analyte cancer screening test. By contrast, TCR analysis did not provide additional sensitivity for stage II-IV disease. Conclusion: We demonstrate detection of a significant fraction of early lung cancer cases from blood by analyzing the circulating TCR repertoire and show that this signal is complementary to established analytes such as proteins and ctDNA. A similar study for the detection of early colorectal cancer and advanced adenomas is under way. Citation Format: Yilong Li, Michelle Nahas, Dennis Stephens, Kate Froburg, Emma Hintz, Devin Champagne, Amaneet Lochab, Markus Brown, Jasper Braun, María Antonia Fortuño, María-del-Mar Ocón, Andrea Pasquier, Inés María Luque, Christopher W. Seder, Jeffrey A. Borgia, Luis Miguel Seijo, Luis M Montuenga, Roman Yelensky. Circulating T-cell receptor repertoire analysis improves cancer early detection [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr A075.

Abstract 4137204: Classifying Left Ventricular Hypertrophy from ECG in Overall Population and Bundle Branch Blocks: Machine Learning Models are Superior to Published ECG Criteria

Background: Published ECG criteria for left ventricular hypertrophy (LVH) have low diagnostic yield, esp. with right/left bundle branch block (RBBB/LBBB). Machine learning (ML) may improve LVH classification. Hypothesis: ML performs better than published LVH criteria to classify LVH from ECG in overall population, and subgroups of RBBB/LBBB. Methods: In addition to standard ECG features like rates, intervals, axis etc., R, S and overall QRS amplitudes, and voltage-time-integrals were extracted from standard 12 leads and reconstructed vectorcardiographic X, Y, Z lead representative beats. Further, a variational autoencoder was used to extract latent encodings from the representative beats. ML algorithms (logistic regression, random forest, light gradient boosted machine, multilayered perceptron, residual network, and convolutional neural network) were trained to predict LVH (LV mass index women &gt;95, men &gt;115 g/m 2 ) from ECG features on a training set of adult ECGs and echocardiogram within 45 days of each other (general population N=240,476, RBBB n=22,110, LBBB n=12,563). We obtained ROC AUCs in a separate validation set for LVH classification by (a) individual ECG variables, (b) published LVH criteria and (c) ML models. Results: As shown in Tables 1-2 and Figure 1 , for LVH classification among the validation sets, AUCs were highest for ML models, intermediate for univariable models and least for published criteria (best ML model vs. best criteria: general population women 0.77 vs. 0.63, men 0.78 vs. 0.63; RBBB women 0.70 vs. 0.57, men 0.77 vs. 0.57; LBBB women 0.71 vs. 0.55, men 0.72 vs. 0.54). Conclusions: ML models perform better than published LVH criteria at classifying LVH from ECG in overall population, RBBB and LBBB.

Exchangeable Copper Excess and Zinc Deficiency in the Serum of Patients with Colorectal Cancer.

Copper (Cu) and Zinc (Zn) are altered in colorectal cancer (CRC) but their association with the clinical classification of the tumor has not been fully explored. To examine the association of Cu and Zn homeostasis in the onset and severity of CRC, we performed an exploratory case-control study comparing the serum levels for Cu, the exchangeable component of Cu in serum (CuExc), Zn, the ratio between them (CuExc:Zn), ceruloplasmin [Cp, concentration (iCp) and its activity (eCp), Cp specific activity (eCp:iCp)], and the Cu:Cp, assessed in 31 consecutive CRC patients before surgical resection to those obtained from 37 healthy controls (CTRL). Additionally, we correlated the analyte levels with the indices of the pathological tumor, node, and metastasis (TNM) staging, namely tumor (T), node (N), and metastasis (M), evaluated at the histopathological examination. We found that Cu, CuExc, CuExc:Zn, iCp, eCp, eCp:iCp, and Cu:Cp ratios increased while Zn decreased in CRC patients. In addition, correlation analyses showed that CuExc and Zn levels confirmed the CRC diagnosis. Specifically, CuExc:Zn further increased the discrimination between the individuals of the two groups, providing an area under the curve (ROC AUC) = 0.94. Elevated CuExc was the strongest factor associated with CRC resulting in 15-fold increased odds. These data were confirmed through a multivariable regression model revealing an effect of Zn and CuExc on the CRC risk, with the CuExc resulting in 11-fold increased odds of having the disease. We also found that most of the Cu biological variables analyzed were associated with T, while the CuExc was associated with M. The current pilot study demonstrates that excess labile Cu pool, Zn deficiency, and even further their combination in the CuExc:Zn provide information about CRC in terms of diagnosis, risk of having CRC, and CRC disease stage.