False Positive Research Articles

Advancing AML tactical approaches with data analytics: Transformative strategies for improving regulatory compliance in banks

The growing complexity of financial crimes necessitates advanced Anti-Money Laundering (AML) strategies that leverage data analytics to improve regulatory compliance in banks. As traditional AML methods face challenges in detecting sophisticated money laundering schemes, data analytics offers transformative solutions by enabling real-time monitoring, enhanced risk detection, and predictive analysis. This review explores the integration of data analytics in AML systems and its impact on regulatory compliance, focusing on strategies that banks can adopt to mitigate risks and adhere to evolving regulations. Data analytics empowers financial institutions to analyze vast amounts of transactional data, identifying suspicious patterns and anomalies with greater precision. Machine learning algorithms and artificial intelligence (AI) further enhance these capabilities by automating risk assessments, reducing false positives, and improving decision-making processes. Through predictive analytics, banks can anticipate emerging threats, adapting their AML strategies proactively to counter new money laundering techniques. A key advantage of data-driven AML approaches is the ability to streamline compliance processes. By automating Know Your Customer (KYC) procedures and cross-referencing data from multiple sources, banks can efficiently verify customer identities and monitor for unusual behavior. Additionally, the adoption of data analytics improves reporting accuracy, ensuring compliance with stringent regulatory frameworks such as the Financial Action Task Force (FATF) and the Bank Secrecy Act (BSA). This review highlights the transformative role of data analytics in enhancing AML efforts, emphasizing the importance of real-time data integration, predictive modeling, and automation. The shift from reactive to proactive AML approaches not only strengthens regulatory compliance but also fosters a culture of vigilance and risk management within banks. As financial institutions continue to embrace digital transformation, leveraging data analytics for AML will be crucial in combating financial crimes and maintaining compliance in an increasingly complex regulatory environment. Keywords: Anti-Money Laundering (AML), Data Analytics, Regulatory Compliance, Banks, Financial Crime, Machine Learning, Artificial Intelligence, Know Your Customer (KYC), Predictive Analytics, Risk Management.

Evaluating the Modified American Academy of Pediatrics Screening Algorithm for Critical Congenital Heart Disease.

In 2018, an expert panel recommended two key modifications to the most used algorithm for screening neonates for critical congenital heart disease (CCHD). Our aim was to evaluate the outcomes of the modified algorithm compared with those of the original algorithm in a real-world setting. We compared the performance characteristics of an original CCHD algorithm used to screen term neonates at a large hospital system between October 26, 2018, and October 15, 2020, and the recommended modified algorithm used between October 15, 2020, and June 30, 2022. We calculated sensitivity, specificity, false positive rate, proportion of false positives with non-CCHD illness, and error rates of test administration and interpretation for each algorithm. Sensitivity was not significantly different between the modified algorithm compared with the original algorithm (40.00 vs. 12.50%, p ≥0.99), but specificity was lower (and hence false positive rate was higher) in the modified algorithm (99.91 vs. 99.98%, p < 0.001). Despite a higher false positive rate in the modified algorithm, the proportion of false positives with significant non-CCHD illness was similar (36.47 vs. 28.57%, p ≥0.99), a finding that translated to an increase in the number of cases of significant non-CCHD illness detected (11 cases out of 32,178 screens vs. 2 cases out of 32,984 screens). Error rates of test administration and interpretation were similar between the two algorithms. In this limited study, the modified algorithm for CCHD screening using pulse oximetry had a higher false positive rate than that of the original American Academy of Pediatrics algorithm. However, this higher rate led to an increased overall number of cases detected of significant non-CCHD illness. · Experts recommend two changes to the American Academy of Pediatrics-endorsed CCHD screen.. · This study evaluates the new algorithm for screening.. · The new algorithm detects at least as many cases as the original.. · There was no significant difference in test sensitivity for CCHD.. · The new algorithm has a statistically higher false positive rate..

Reliable molecular detection of small hive beetles

Invasive species require adequate reliable detection methods to mitigate their further spread and impact. However, the reliability of molecular detection methods is often hampered by both false positives (Error type I) and false negatives (Error type II). At present, the reliability of the four published molecular detection methods for small hive beetles (SHB), Aethina tumida, has not been rigorously evaluated considering their extensive genetic diversity. Here, we performed intra- and interlaboratory comparisons of the four available methods using SHB samples representing 78 regions from 27 countries on five continents, beetles from the same genus (Aethina concolor, A. inconspicua, A. flavicollis and A. major), as well as western honey bees, Apis mellifera, and ectoparasitic mites Varroa destructor. The data show that the Idrissou et al. (2018) and Li et al. (2018) methods avoid both false positives and false negatives probably due to lower sensitivity to nucleotide mismatches on the primer and probe’s target sequences. Further, the Li et al. (2018) method can be considered more sensitive because the fluorescent amplification curve crosses the threshold at lower Cq values compared to the Idrissou et al. (2018) one. In light of our data, the Li et al. (2018) method is the most reliable molecular diagnostic tool for SHB. We therefore recommend using this method as it will contribute to management efforts of this invasive species.

Machine Learning in Cybersecurity: A Comprehensive Analysis of Intrusion Detection Systems

Cyber threats are getting more complex and larger, and traditional security measures can't keep up with the sophisticated attacks. Machine Learning (ML) has emerged as a powerful tool in enhancing cybersecurity, particularly in the development of Intrusion Detection Systems (IDS). a comprehensive analysis of the role of machine learning in intrusion detection, focusing on its ability to detect and respond to both known and unknown threats in real-time. various machine learning techniques, including supervised, unsupervised, and deep learning models, and evaluates their effectiveness in identifying anomalies and preventing security breaches. Key performance metrics such as accuracy, false positive rates, and scalability are analyzed, with a focus on the advantages and limitations of each approach. Additionally, the paper explores the challenges associated with implementing ML-based IDS, including data quality, adversarial attacks, and computational requirements. Through case studies and real-world examples, this paper highlights how machine learning is revolutionizing intrusion detection and presents insights into future advancements that will further strengthen cybersecurity defenses.

Value of Functional Protein S Activity Assays for Detection of Type II Protein S Deficiency

Abstract Background Protein S deficiency is a hereditary thrombophilia which occurs in three forms: Type I (quantitative), type II (qualitative) and type III (qualitative due to a gain of function). Literature suggests that of these, type II is the rarest and most difficult to diagnose. Type II protein S deficiency is only detectable by assaying protein S activity with a functional, clot-based test. However, this test is susceptible to interference and has a lower specificity when compared to antigen assays used to detect free and total protein S levels. Subsequently, the Choosing Wisely Guidelines released by the ABIM Foundation recommends against using protein S activity assays entirely. Instead, they suggest using only free protein S antigen assays. Although the rationale is to avoid false positive results from the protein S activity assay, it ultimately may result in a failure to detect type II protein S deficiency. Objective This study, performed at a tertiary medical center in the Midwestern region of the United States, aims to determine the proportion of false positives in functional assays as well as the percent of true type II protein S deficiency cases within this testing population. Methods All cases with abnormal protein S function within the institution were evaluated from July 2018 through July 2023 and classified as acquired or hereditary deficiencies. Cases were classified as acquired if there was evidence of liver disease, vitamin K deficiency, ongoing thrombosis, consumption, disseminated intravascular coagulopathy, or pregnancy. Possible hereditary deficiency cases were classified as type I, type II and type III: Type I if all assays were low, type III if both active and free protein S assays were low with normal total protein S, and type II if only protein S activity was low. The cutoff defined for normal total and free protein S levels was greater than 70% activity, and the low protein S activity cutoff was defined as less than 50%. Patients with low protein S activity but normal immunologic protein S and an elevated factor VIII were considered falsely low values. Chart review will be performed for remaining cases with low protein S activity to rule out confounders such as vitamin K deficiency, warfarin use, acute phase reaction, or pregnancy. Results From a total of 8805 panels performed to assess for possible thrombophilia, 1580 (18%) were found to have decreased protein S activity. Of these cases, 599 (38%) were considered falsely decreased due to elevated factor VIII levels. 322 cases (3.7%) were considered possible hereditary deficiencies: 217 type III (2.5%), 27 type II (0.3%), and 80 type I (0.9%). Discussion Our data indicates the prevalence of subtypes may be differently distributed than previously believed, which may call for current guidelines to be reevaluated.

Seamless, rapid, and accurate analyses of outbreak genomic data using split k-mer analysis.

Sequence variation observed in populations of pathogens can be used for important public health and evolutionary genomic analyses, especially outbreak analysis and transmission reconstruction. Identifying this variation is typically achieved by aligning sequence reads to a reference genome, but this approach is susceptible to reference biases and requires careful filtering of called genotypes. There is a need for tools that can process this growing volume of bacterial genome data, providing rapid results, but that remain simple so they can be used without highly trained bioinformaticians, expensive data analysis, and long-term storage and processing of large files. Here we describe split k-mer analysis (SKA2), a method that supports both reference-free and reference-based mapping to quickly and accurately genotype populations of bacteria using sequencing reads or genome assemblies. SKA2 is highly accurate for closely related samples, and in outbreak simulations, we show superior variant recall compared with reference-based methods, with no false positives. SKA2 can also accurately map variants to a reference and be used with recombination detection methods to rapidly reconstruct vertical evolutionary history. SKA2 is many times faster than comparable methods and can be used to add new genomes to an existing call set, allowing sequential use without the need to reanalyze entire collections. With an inherent absence of reference bias, high accuracy, and a robust implementation, SKA2 has the potential to become the tool of choice for genotyping bacteria. SKA2 is implemented in Rust and is freely available as open-source software.

Improved identification of network anomalies through optimal CURE clustering

Abstract In this paper, we propose an advanced network anomaly behavior identification framework to overcome the constraints inherent in conventional rule- or signature-based approaches, which often struggle with emerging and previously unknown threats. Central to our framework is an Enhanced CURE (Cluster Updating and REfining) clustering algorithm, meticulously tailored and refined to incorporate a density-based methodology. This enhancement enables the algorithm to discern subtle shifts in network anomaly patterns with heightened precision. The implementation workflow commences with the application of the optimized CURE algorithm to analyze network data, followed by the deployment of a sophisticated anomaly degree ranking mechanism. This mechanism, through meticulous calculation of individual data points' anomaly degrees and subsequent ranking, effectively isolates those deviating significantly from standard behavioral norms, incorporating a strategic threshold to filter out false positives. To validate our methodology's efficacy and its superiority over existing techniques, experiments were conducted utilizing a substantial real-world network dataset. These tests affirm not only a marked increase in the accuracy of abnormal behavior recognition and a reduction in computational intricacy but also demonstrate the adaptability across diverse network ecosystems. Our approach has proven successful in pinpointing a wide array of network anomalies, encompassing malicious cyberattacks, fraudulent activities, unauthorized intrusions, and breaches of security protocols, thereby highlighting its comprehensive capability in bolstering network defense strategies. Despite the notable advancements and successful identification of various network anomalies, our framework currently lacks integration with real-time learning capabilities, limiting its immediate responsiveness to rapidly evolving attack patterns and necessitating ongoing research for dynamic updates and adaptive learning mechanisms.

Don’t omit a urine culture when a urinalysis is suspicious for asymptomatic bacteriuria during pregnancy

Abstract Introduction/Objective Investigation of a clinical problem using steps of evidence-based practice (EBP) led to the development of a recommendation submitted for Choosing Wisely consideration. The prevalence of asymptomatic bacteriuria (ASB) in pregnancy is between 2-15%. ASB may be a concern in pregnancy because of its capacity for association with adverse outcomes, namely preterm delivery, low birthweight, and possible progression to pyelonephritis. Screening and treatment may reduce the risks of adverse outcomes, but data is not current. A point-of-care urine dipstick is low cost and commonly used at routine obstetrician clinic visits to screen for ASB. The urinalysis dipstick is associated with high false-positive and false-negative rates as compared to the gold standard urine culture. Some providers make the decision to treat ASB given a positive urinalysis screen without ordering a follow-up culture. Such practice could lead to overtreatment of the patient and poor antibiotic stewardship. Research question Is a urinalysis without culture sufficient for diagnosis and treatment of ASB during pregnancy? Methods/Case Report Laboratory Medicine Best Practices A-6 steps of Ask, Acquire, Appraise, Apply, Analyze, and Assess were utilized to develop an appropriate and relevant clinical question, collect literature, determine evidence value, assign ratings, implement into practice, and evaluate outcomes. During appraisal, six selected articles were graded according to study rating, practice, measures, and results, further including the final rating and effect size rating. Consistency and overall strength of the evidence were analyzed. Results (if a Case Study enter NA) Each evidence study demonstrated final ratings between 8 and 10, a rating of “Good,” and an effect size rating of “Substantial.” Consistency leads to a “High” overall strength among the body of evidence. Conclusion Obtain a urine culture if routine prenatal point-of-care dipstick urinalyses have findings consistent with asymptomatic bacteriuria. Urinalyses are limited as screening due to low sensitivity and specificity for bacteriuria. Positive urinalysis dipstick results should be neither the sole basis for diagnosis nor decision to treat.

Deep Learning Approach for Fetal Health Prediction

Monitoring the foetus's health is crucial during pregnancy to avoid complications that may worsen the course of pregnancy and delivery. Cardiotocography (CTG) is a tool that provides complex information by monitoring the foetus's heart rate signal. Obstetricians visually interpret these signals to predict potential risks and draw clinical inferences. The interpretation, however, relies on the expertise of the obstetrician, leading to a significant false positive rate. Thus, the study uses deep learning techniques to effectively identify foetal health states as 'Normal', 'Suspect' and 'Pathological'. The dataset used is CTG data drawn from the Kaggle repository. The optimal subset of features is obtained by comparing traditional feature selection and meta-heuristic-based feature selection techniques. Deep learning algorithms, namely, Convolution Neural Networks, Artificial Neural Networks, and Radial Basis Function Networks, are applied to train multi-class classification models that predict foetal health status. The models are then evaluated using performance metrics.

Modular Point-of-Need Tropane Alkaloid Detection at Regulatory Levels: Combining Solid-Liquid Extraction from Buckwheat with a Paper-Immobilized Liquid-Phase Microextraction and Immuno-Detection in Interconnectable 3D-Printed Devices.

Contamination with tropane alkaloids in cereals is expected to increase globally. However, current identification tools (e.g., liquid chromatography-mass spectrometry) for tropane alkaloids are time-consuming and expensive. Furthermore, their miniaturized alternatives lack sensitivity and robustness. Therefore, there is a pressing need for inexpensive and effective screening methods. Here, an on-site applicable modular workflow for tropane alkaloid detection in buckwheat is presented. The modular workflow combines paper microfluidics and interconnectable 3D-printed sample preparation tools and was evaluated for different tropane alkaloids, including atropine and scopolamine. Furthermore, integration with an indirect competitive lateral flow immunoassay (icLFIA) for atropine detection at relevant levels was demonstrated. In the modular workflow, to minimize matrix coextraction, tropane alkaloids were extracted from the milled buckwheat cereals by a mixture of alkaline aqueous and immiscible organic solvents (extraction recoveries: 66-79%). The tropane alkaloids were subsequently concentrated with a newly developed paper-immobilized liquid-phase microextraction (PI-LPME, extraction recoveries: 34-60%, concentration factor to immobilized solution in paper: 60-108×). After the PI-LPME, with an integrated 3D-printed setup, the tropane alkaloids were directly eluted (elution recoveries: 83-93%) and detected with the icLFIA. Digital read-out of the icLFIA, by employing a hand-held reader, enabled semiquantification of atropine (IC50 = 0.56 ng mL-1 in standard solutions). The modular workflow was validated by analyzing 24 blank and spiked buckwheat cereal samples with 5 and 10 μg kg-1 atropine. A cutoff value was established with an estimated false negative rate of 1% and estimated false positive rate of 0.68%. Therefore, the modular workflow can aid in fast, inexpensive, and on-site atropine detection by nonexperts, and when integrated with a scopolamine-specific icLFIA expanded toward scopolamine detection. Moreover, the developed sample extraction and concentration method (PI-LPME) is suitable for the analysis of many other compounds with pH-dependent polarity.

SiT-YOLOv9: An Efficient Algorithm for Learning Behavior Detection in the Home Environment

In the context of home-based learning, accurate identification of learning behaviors is essential for enhancing post-classroom learning efficiency. However, due to background interference and computational constraints in the TinyML terminal within home environments, CNN-based algorithms are susceptible to reduced performance and accuracy, leading to an increased false positive rate. To address this challenge, we propose a lightweight detection model called SiT-YOLOv9, which integrates MODNet, image enhancement, and other modules into the YOLOv9 model while also implementing moderate network pruning to effectively mitigate issues related to image noise and training sample computational power. Evaluation of the SiTBehaviors video dataset demonstrates that the SiT YOLOv9 model achieves outstanding performance with a recognition accuracy of 0.948 (mAP50) at a high processing speed of 90.9 frames per second. When compared with original models such as YOLOv8, YOLOv10, and RT-DERT, our proposed model exhibits superior recognition accuracy of 0.948 mAP, and a processing speed of 0.2 ms.

TSAE-UNet: A Novel Network for Multi-Scene and Multi-Temporal Water Body Detection Based on Spatiotemporal Feature Extraction

The application of remote sensing technology in water body detection has become increasingly widespread, offering significant value for environmental monitoring, hydrological research, and disaster early warning. However, the existing methods face challenges in multi-scene and multi-temporal water body detection, including the diverse variations in water body shapes and sizes that complicate detection; the complexity of land cover types, which easily leads to false positives and missed detections; the high cost of acquiring high-resolution images, limiting long-term applications; and the lack of effective handling of multi-temporal data, making it difficult to capture the dynamic changes in water bodies. To address these challenges, this study proposes a novel network for multi-scene and multi-temporal water body detection based on spatiotemporal feature extraction, named TSAE-UNet. TSAE-UNet integrates convolutional neural networks (CNN), depthwise separable convolutions, ConvLSTM, and attention mechanisms, significantly improving the accuracy and robustness of water body detection by capturing multi-scale features and establishing long-term dependencies. The Otsu method was employed to quickly process Sentinel-1A and Sentinel-2 images, generating a high-quality training dataset. In the first experiment, five rectangular areas of approximately 37.5 km2 each were selected to validate the water body detection performance of the TSAE-UNet model across different scenes. The second experiment focused on Jining City, Shandong Province, China, analyzing the monthly water body changes from 2020 to 2022 and the quarterly changes in 2022. The experimental results demonstrate that TSAE-UNet excels in multi-scene and long-term water body detection, achieving a precision of 0.989, a recall of 0.983, an F1 score of 0.986, and an IoU of 0.974, significantly outperforming FCN, PSPNet, DeepLabV3+, ADCNN, and MECNet.

Unsupervised Security Threats Identification for Heterogeneous Events

As cyberattacks targeting industrial control systems continue to evolve, the development of sophisticated technologies to detect these security threats becomes increasingly essential. In addition, it is necessary to update adversarial information constantly. However, this process is complicated by the deployment of heterogeneous equipment, which increases the number of indicators and characteristics that must be analyzed by security administrators. Furthermore, security operation centers often struggle to respond promptly to adversaries because of the high number of false alerts caused by unreliable system labels. These challenges make it difficult to construct reliable detection systems. To address these issues, we propose a robust unsupervised threat-identification method. Our approach involves applying a preprocessing technique tailored to the various data types pertinent to alerts, followed by classifying unlabeled alerts using an autoencoder (AE) model. Despite the presence of numerous false positives, we verified that the proposed model could effectively distinguish between different attack types and identify their relationships with only one round of training in homogeneous and heterogeneous environments within industrial control systems. Moreover, our model can filter and display data classified as actual attacks and generate relational tables.

Ejection of Real-Time Malicious Intrusion and Attacks in IoT Empowered Infrastructure

The project recognizes the pervasive threat of computer viruses, malware, and hostile attacks on computer networks, highlighting the critical role of intrusion detection as a proactive defense technology. The project introduces a novel approach based on deep learning to identify and mitigate cybersecurity vulnerabilities and breaches in IoT-driven cyber-physical systems, aiming for enhanced security measures. The project's objective is to elevate intrusion detection beyond the limitations of traditional systems by addressing issues like accuracy, detection effectiveness, and reducing false positives. This emphasizes the advancement and innovation in cybersecurity. To achieve the project's goals, the method employs a generative adversarial network, a cutting-edge deep learning technique. Additionally, it distinguishes itself by contrasting unsupervised and deep learning-based discriminative approaches, showcasing a comprehensive and effective approach to cybersecurity. In our project, we successfully implemented an ensemble method to boost predictive accuracy by integrating multiple individual models. Particularly noteworthy is the inclusion of a hybrid architecture, combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM), denoted as CNN+LSTM. This hybrid model achieved an impressive accuracy of 99% when applied to the KDD-Cup dataset, underscoring the efficacy of our ensemble technique for intrusion detection in IoT- based cybersecurity infrastructures.

A Tale of Two Cases: Acquired Hemophilia A and/or Lupus Anticoagulant?

Abstract Distinguishing between acquired hemophilia A (AHA) and lupus anticoagulant (LA) is essential for effective clinical management. This can be challenging due to similarities in laboratory tests and potential interference between their diagnostic tests. Here we present two complex cases. Case 1 is a 69-year-old female with PNH presenting with a subdural hematoma. Prolonged aPTT with no correction on mixing raised clinical suspicion of AHA. Factor VIII activity was low (2%) across three dilutions in the factor assay. A Bethesda assay revealed a low factor VIII inhibitor (1%). LA was negative by DRVVT. Antiphospholipid antibodies (Cardiolipin and Beta-2-Glycoprotein 1) were also negative. Treatment with prednisone and intravenous immunoglobulin was initiated with presumptive diagnosis of AHA. Subsequent samples to check other factor levels uncovered low Factor IX as well. Higher dilutions in factor assays exhibited non-parallelism in Factor IX and VIII, suggesting a non-specific inhibitor. Chromogenic Factor VIII was within the normal range. Another LA test, Staclot-LA was positive. Overall, this case was consistent with LA rather than AHA. The initial ‘positive’ Bethesda assay was attributed to interference by LA. Case 2 is a 20-year-old female with no significant medical history who was admitted elsewhere for a complicated C-section with intraabdominal hemorrhage. She had prolonged aPTT and positive LA by platelet neutralization test there. Upon transfer, the initial assessment confirmed prolonged aPTT and a time-dependent inhibitor pattern in mixing studies. Factor studies revealed significantly diminished factor VIII activity (&amp;lt;1%) using both one-stage clotting and chromogenic assays. A strong factor VIII inhibitor was identified through chromogenic Bethesda assay (126 BU). Non-parallelism was observed in other factors (Factor IX, XI, XII). LA was negative by DRVVT but positive by Staclot-LA. Cardiolipin and beta2-glycoprotein-1 antibodies were negative. The diagnosis of AHA was evident. But is there concurrent LA present? Both Staclot-LA and platelet neutralization tests are PTT-based and susceptible to interference from factor VIII inhibitor. The patient underwent treatment, and factor VIII inhibitor disappeared in 3 months. Staclot-LA also turned negative, suggesting that the initially ‘positive’ LA test was likely a result of interference. In both cases, adopting a holistic approach to differentiate between AHA and LA is beneficial. Relying solely on individual tests may lead to false interpretations, as LA can inhibit FVIII activity assays and produce false positives in the Bethesda assay. It is advisable to include higher dilutions in factor assays and/or chromogenic assays to mitigate interference. The presence of factor VIII inhibitor can also result in false-positive LA results such as Staclot-LA assay. For more reliable results, DRVVT and antibody tests are preferable. If the aPTT-based assay is the sole positive LA test, the presence of LA is suggested when it occurs before and/or after the presence of a factor inhibitor.

The standard posture is a myth: a scoping review.

The standard posture described in Kendall's manual is commonly used for postural assessment. However, no bibliographic reference was provided to support its use. To identify the original source and the procedure followed for the design of that posture and to compare it with current literature on the subject. In accordance with the PRISMA Extension for Scoping Reviews recommendations, PubMed and Scopus were searched using the terms "standing posture", "plum line," and "gravity line". Publications in English, French, German, or Spanish that referred to posture in adults without pathology were included. Six articles and 3 books were included in the final analysis. An identical posture to that described in Kendall's manual was identified in an early 19th-century work carried out with the unrealistic objective of maintaining static bipedal standing without muscular support, and including several anatomical misconceptions. Furthermore, the "ideal alignment" described in Kendall's manual does not correspond to the actual line of gravity, the comfortable posture, or natural postural compensations due to age, gender, or race. The utilization of this standard to ascertain postural deficiencies is not supported by current evidence and may result in numerous false positives, particularly in the elderly.

Ultrasensitive detection of circulating tumor DNA using a CRISPR/Cas9 nickase-driven 3D DNA walker based on a COF-AuNPs sensing platform.

Aelectrochemical biosensor was designedutilizing a CRISPR Cas9n-driven DNA walker combined with gold-nanosphere-like covalent organic frameworks (COFs-AuNPs) to detect breast cancer markers (PIK3CA E545K ctDNA). The DNA walker probe is activated only in the presence of circulating tumor deoxyribonucleic acid(ctDNA), binding to a support probe to form a double strand that is then specifically cleaved by the Cas9n/sgRNA complex. This cleavage produces numerous DNA fragments for signal amplification. The COF-AuNPs as electrode materials facilitate electronic transfer and provide additional active sites for the immobilization of nucleic acid probes. This setup achieves a detection limit of 1.76 aM, demonstrating high sensitivity. Additionally, Cas9n improves the specificity of the sensor, accurately distinguishing a pair of base-mismatched sequences, and reducing the occurrence of false positives. Overall, the sensor exhibits excellent selectivity, reproducibility, and potential for early diagnosis of breast cancer.

Integrating Whole Slide Imaging and Artificial Intelligence Provides Higher Accuracy in Thyroid Cytology

Abstract Introduction/Objective The aim of the study is to evaluate the potential advantages of integrating Whole Slide Imaging (WSI) and Artificial Intelligence (AI) to increase efficiency and diagnostic accuracy in thyroid cytology, especially in indeterminate lesions. Methods/Case Report A retrospective database search from 2000-2023 was performed, and the concordance with the pathologists was evaluated based on the manual or AI-based Region of Interest (ROI) identification, the best- suited AI, algorithm design, color enhancement, and combination of AI and Electronic Medical Records (EMR). Results (if a Case Study enter NA) The cohort included 3285 WSI and 121,384 ROI. In 109 WSI with manually diagnosed ROI, Deep learning successfully predicted benign and malignant cases, including 11 indeterminate cases, with zero false positives or negatives for 35% of cases. In the second study with 908 WSI, 4494 ROI were identified with Convolutional Neural Network (CNN). A higher area under the curve (AUC) resulted when EMR and Machine Learning algorithm (MLA) were integrated: EMR: 0.931, MLA: 0932, EMR+ MLA; 0.962, Sensitivity: 92.0%, Specificity 90.5%. In the next study with 145 WSI, the CNN-identified ROI was presented to the same pathologist after 117 days. Perfect concordance was seen between the two methods (K score 0.924, time to diagnosis;81.59 Second). In 360 WSI, YOLOV4 algorithm was trained to detect the ROI, and detection and classification rates were used to evaluate the accuracy. It had a precision of 81.84% (3.16% more than the single detection networks of 78.68%). In the last study with 964 WSI, color augmentation improved the AUC for indeterminate cases from 88.3% to 96.2%. Conclusion CNN is a better-suited AI for whole slide imaging due to its compatibility with image data and ability to automatically detect important features without human supervision. It also removes the component of subjective bias in interpretation. Color augmentation and integration of EMR can overcome the unique challenges in thyroid indeterminate cytology.

Machine learning to classify left ventricular hypertrophy using ECG feature extraction by variational autoencoder.

Traditional ECG criteria for left ventricular hypertrophy (LVH) have low diagnostic yield. Machine learning (ML) can improve ECG classification. ECG summary features (rate, intervals, axis), R-wave, S-wave and overall-QRS amplitudes, and QRS/QRST voltage-time integrals (VTIs) were extracted from 12-lead, vectorcardiographic X-Y-Z-lead, and root-mean-square (3D) representative-beat ECGs. Latent features were extracted by variational autoencoder from X-Y-Z and 3D representative-beat ECGs. Logistic regression, random forest, light gradient boosted machine (LGBM), residual network (ResNet) and multilayer perceptron network (MLP) models using ECG features and sex, and a convolutional neural network (CNN) using ECG signals, were trained to predict LVH (left ventricular mass indexed in women >95 g/m², men >115 g/m²) on 225,333 adult ECG-echocardiogram (within 45 days) pairs. AUROCs for LVH classification were obtained in a separate test set for individual ECG variables, traditional criteria and ML models. In the test set (n=25,263), AUROC for LVH classification was higher for ML models using ECG features (LGBM 0.790, MLP 0.789, ResNet 0.788) as compared to the best individual variable (VTI QRS-3D 0.677), the best traditional criterion (Cornell voltage-duration product 0.647) and CNN using ECG signal (0.767). Among patients without LVH who had a follow-up echocardiogram >1 (closest to 5) years later, LGBM false positives, compared to true negatives, had a 2.63 (95% CI 2.01, 3.45)-fold higher risk for developing LVH (p<0.0001). ML models are superior to traditional ECG criteria to classify-and predict future-LVH. Models trained on extracted ECG features, including variational autoencoder latent variables, outperformed CNN directly trained on ECG signal.

A hybrid LSTM random forest model with grey wolf optimization for enhanced detection of multiple bearing faults

Bearing degradation is the primary cause of electrical machine failures, making reliable condition monitoring essential to prevent breakdowns. This paper presents a novel hybrid model for the detection of multiple faults in bearings, combining Long Short-Term Memory (LSTM) networks with random forest (RF) classifiers, further enhanced by the Grey Wolf Optimization (GWO) algorithm. The proposed approach is structured in three stages: first, time and frequency domain features are manually extracted from vibration signals; second, these features are processed by a dual-layer LSTM network, which is specifically designed to capture complex temporal relationships within the data; finally, the GWO algorithm is employed to optimize feature selection from the LSTM outputs, feeding the most relevant features into the RF classifier for fault classification. The model was rigorously evaluated using a dataset comprising six distinct bearing health conditions: healthy, outer race fault, ball fault, inner race fault, compounded fault, and generalized degradation. The hybrid LSTM-RF-GWO model achieved a remarkable classification accuracy of 98.97%, significantly outperforming standalone models such as LSTM (93.56%) and RF (98.44%). Furthermore, the inclusion of GWO led to an additional accuracy improvement of 0.39% compared to the hybrid LSTM-RF model without optimization. Other performance metrics, including precision, kappa coefficient, false negative rate (FNR), and false positive rate (FPR), were also improved, with precision reaching 99.28% and the kappa coefficient achieving 99.13%. The FNR and FPR were reduced to 0.0071 and 0.0015, respectively, underscoring the model’s effectiveness in minimizing misclassifications. The experimental results demonstrate that the proposed hybrid LSTM-RF-GWO framework not only enhances fault detection accuracy but also provides a robust solution for distinguishing between closely related fault conditions, making it a valuable tool for predictive maintenance in industrial applications.