High False-negative Rate Research Articles

Multi-classification in-vehicle intrusion detection system using packet- and sequence-level characteristics from time-embedded transformer with autoencoder

In recent years, the increased application of controller area network (CAN) protocols has made it the de facto standard for communication between electronic control units (ECUs) in the automotive and transportation fields. This widely used protocol was designed as a reliable and straightforward broadcast-based protocol that connects ECUs without considering security concerns such as node authentication or traffic encryption. Despite its efficiency, this tradeoff makes the CAN bus vulnerable to attacks. Implementing intrusion detection systems (IDSs) based on machine learning (ML) can address these security challenges effectively. However, existing ML-based IDSs have limited classification capabilities, lack adaptability and time sensitivity, incomprehensive analysis, and produce high false-negative rates (FNR), while attack schemes are becoming increasingly complex, resulting in insufficient capability of intrusion detection in real-time and insufficient ability to offer reliable protection. Therefore, our study proposes a novel in-vehicle IDS for multiclass classification using both packet- and sequence-level characteristics extracted from an autoencoder and a variant transformer (Time-embedded Transformer) with an improved position encoding mechanism, which analyses CAN traffic in-depth from various perspectives to overcome the existing challenges above. Both standard (Car-Hacking) and advanced (ROAD) datasets are used to evaluate the capabilities of our proposed IDS. The evaluation results demonstrated 100 % detection accuracy and 0 % FNR for both the Car-Hacking and ROAD Masquerade datasets, which also peaked at the highest F1 score for the ROAD Fabrication dataset, emphasizing superior intrusion detection to minimize FNR of the proposed model with high adaptability through its multi-dimensional analysis at packet- and sequence-level.

Usability of the Japanese Late-Stage Elderly Questionnaire for screening major depression.

Older adults with major depression are at risk of frailty and long-term care needs. Consequently, screening for major depression is imperative to prevent such risks. In Japan, the Late-Stage Elderly Questionnaire was developed to evaluate older adults' holistic health, including mental well-being. It comprises one specific question to gauge life satisfaction, but the effectiveness of this question to screen for major depression remains unclear. Therefore, we aimed to assess the usability of this question to screen for major depression. This retrospective cohort study used a large, commercially available claims database in Japan. Participants were older adults aged ≥75 years who completed the Late-Stage Elderly Questionnaire and were classified with and without new major depression within 1 year. We evaluated the questionnaire's ability to screen for major depression using C-statistics, developing three models to assess the cut-off value based on responses to the life satisfaction question ('Satisfied', 'Somewhat satisfied', 'Somewhat unsatisfied', or 'Unsatisfied'), estimating the sensitivity and specificity of each model. Among 11 117 older adults, 77 newly experienced major depression within 1 year. The C-statistic for screening major depression was 0.587. The model setting the cut-off between 'Somewhat unsatisfied' and 'Unsatisfied' the demonstrated lowest sensitivity and highest specificity, while the model setting the cut-off between 'Satisfied' and 'Somewhat satisfied' demonstrated highest sensitivity and lowest specificity. Our results suggest that due to its poor screening ability and high rate of false negatives, the question assessing life satisfaction in the Late-Stage Elderly Questionnaire may not be useful for screening major depression in older adults and may require modification.

Computer-aided detection of prostate cancer in early stages using multi-parameter MRI: A promising approach for early diagnosis.

Transrectal ultrasound-guided prostate biopsy is the gold standard diagnostic test for prostate cancer, but it is an invasive examination of non-targeted puncture and has a high false-negative rate. In this study, we aimed to develop a computer-assisted prostate cancer diagnosis method based on multiparametric MRI (mpMRI) images. We retrospectively collected 106 patients who underwent radical prostatectomy after diagnosis with prostate biopsy. mpMRI images, including T2 weighted imaging (T2WI), diffusion weighted imaging (DWI), and dynamic-contrast enhanced (DCE), and were accordingly analyzed. We extracted the region of interest (ROI) about the tumor and benign area on the three sequential MRI axial images at the same level. The ROI data of 433 mpMRI images were obtained, of which 202 were benign and 231 were malignant. Of those, 50 benign and 50 malignant images were used for training, and the 333 images were used for verification. Five main feature groups, including histogram, GLCM, GLGCM, wavelet-based multi-fractional Brownian motion features and Minkowski function features, were extracted from the mpMRI images. The selected characteristic parameters were analyzed by MATLAB software, and three analysis methods with higher accuracy were selected. Through prostate cancer identification based on mpMRI images, we found that the system uses 58 texture features and 3 classification algorithms, including Support Vector Machine (SVM), K-nearest Neighbor (KNN), and Ensemble Learning (EL), performed well. In the T2WI-based classification results, the SVM achieved the optimal accuracy and AUC values of 64.3% and 0.67. In the DCE-based classification results, the SVM achieved the optimal accuracy and AUC values of 72.2% and 0.77. In the DWI-based classification results, the ensemble learning achieved optimal accuracy as well as AUC values of 75.1% and 0.82. In the classification results based on all data combinations, the SVM achieved the optimal accuracy and AUC values of 66.4% and 0.73. The proposed computer-aided diagnosis system provides a good assessment of the diagnosis of the prostate cancer, which may reduce the burden of radiologists and improve the early diagnosis of prostate cancer.

Comparison of three colloidal gold immunoassays and GeneXpert Carba-R for the detection of Klebsiella pneumoniae blaKPC-2 variants.

The increasing use of ceftazidime-avibactam has led to the emergence of a wide range of ceftazidime-avibactam-resistant blaKPC-2 variants. Particularly, the conventional carbapenemase phenotypic assay exhibited a high false-negative rate for KPC-2 variants. In this study, three colloidal gold immunoassays, including the Gold Mountainriver CGI test, Dynamiker CGI test and NG-Test CARBA5, and GeneXpert Carba-R, were used to detect the presence of KPC-2 carbapenemase and its various variants in 42 Klebsiella pneumoniae strains. These strains covered blaKPC-2 (13/42) and 16 other blaKPC-2 variants including blaKPC-12 (1/42), blaKPC-23 (1/42), blaKPC-25 (1/42), blaKPC-33 (6/42), blaKPC-35 (1/42), blaKPC-44 (1/42), blaKPC-71 (1/42), blaKPC-76 (8/42), blaKPC-78 (1/42), blaKPC-79 (1/42), blaKPC-100 (1/42), blaKPC-127 (1/42), blaKPC-128 (1/42), blaKPC-144 (1/42), blaKPC-157 (2/42), and blaKPC-180 (1/42). For KPC-2 strains, all four assays showed 100% negative percentage agreement (NPA) and 100% positive percentage agreement (PPA) with sequencing results. For all 16 KPC-2 variants, GeneXpert Carba-R showed 100% NPA and 100% PPA, and the three colloidal gold immunoassays showed 100% NPA, while the PPAs of the Gold Mountainriver CGI test, Dynamiker CGI test, and NG-Test CARBA5 were 87.5%, 87.5%, and 68.8%, respectively. We also found a correlation between the mutation site in the amino acid of the variants and false-negative results by colloidal gold immunoassays. In conclusion, the GeneXpert Carba-R has been proven to be a reliable method in detecting KPC-2 and its variants, and the colloidal gold immunoassay tests offer a practical and cost-effective approach for their detection. For the sample with a negative result by a colloidal gold immunoassay test but not matching the drug-resistant phenotype, it is recommended to retest using another type of kit or the GeneXpert Carba-R assay, which can significantly improve the accuracy of detection.

A hybrid feature weighted attention based deep learning approach for an intrusion detection system using the random forest algorithm.

Due to the recent advances in the Internet and communication technologies, network systems and data have evolved rapidly. The emergence of new attacks jeopardizes network security and make it really challenging to detect intrusions. Multiple network attacks by an intruder are unavoidable. Our research targets the critical issue of class imbalance in intrusion detection, a reflection of the real-world scenario where legitimate network activities significantly out number malicious ones. This imbalance can adversely affect the learning process of predictive models, often resulting in high false-negative rates, a major concern in Intrusion Detection Systems (IDS). By focusing on datasets with this imbalance, we aim to develop and refine advanced algorithms and techniques, such as anomaly detection, cost-sensitive learning, and oversampling methods, to effectively handle such disparities. The primary goal is to create models that are highly sensitive to intrusions while minimizing false alarms, an essential aspect of effective IDS. This approach is not only practical for real-world applications but also enhances the theoretical understanding of managing class imbalance in machine learning. Our research, by addressing these significant challenges, is positioned to make substantial contributions to cybersecurity, providing valuable insights and applicable solutions in the fight against digital threats and ensuring robustness and relevance in IDS development. An intrusion detection system (IDS) checks network traffic for security, availability, and being non-shared. Despite the efforts of many researchers, contemporary IDSs still need to further improve detection accuracy, reduce false alarms, and detect new intrusions. The mean convolutional layer (MCL), feature-weighted attention (FWA) learning, a bidirectional long short-term memory (BILSTM) network, and the random forest algorithm are all parts of our unique hybrid model called MCL-FWA-BILSTM. The CNN-MCL layer for feature extraction receives data after preprocessing. After convolution, pooling, and flattening phases, feature vectors are obtained. The BI-LSTM and self-attention feature weights are used in the suggested method to mitigate the effects of class imbalance. The attention layer and the BI-LSTM features are concatenated to create mapped features before feeding them to the random forest algorithm for classification. Our methodology and model performance were validated using NSL-KDD and UNSW-NB-15, two widely available IDS datasets. The suggested model's accuracies on binary and multi-class classification tasks using the NSL-KDD dataset are 99.67% and 99.88%, respectively. The model's binary and multi-class classification accuracies on the UNSW-NB15 dataset are 99.56% and 99.45%, respectively. Further, we compared the suggested approach with other previous machine learning and deep learning models and found it to outperform them in detection rate, FPR, and F-score. For both binary and multiclass classifications, the proposed method reduces false positives while increasing the number of true positives. The model proficiently identifies diverse network intrusions on computer networks and accomplishes its intended purpose. The suggested model will be helpful in a variety of network security research fields and applications.

1D-CNN: Classification of normal delivery and cesarean section types using cardiotocography time-series signals

Abstract Cardiotocography (CTG) is considered the gold standard for monitoring fetal heart rate (FHR) during pregnancy and labor to estimate the danger of oxygen deprivation. Visual interpretation of CTG traces is complex and frequently results in high rates of false positives and false negatives, leading to unfavorable and unwanted outcomes such as fetal mortality or needless cesarean surgery. If the data are well-balanced, which is uncommon in medical datasets, machine learning techniques can be helpful in interpretation. This study is designed to determine classification performance under various data balance approaches. We propose a robust methodology for the automated extraction of features that use a deep learning model based on the one-dimensional convolutional neural network (1D-CNN). We used a public database containing 552 intrapartum CTG recordings. Due to the imbalance in the dataset, the experiments were conducted under a variety of conditions such as (i) an unbalanced dataset, (ii) undersampling, (iii) a weighted binary cross-entropy approach, and (iv) oversampling utilizing the synthetic minority oversampling technique (SMOTE). We found an excellent sensitivity (99.80% for the unbalanced dataset, 96.25% for the weighted binary cross-entropy approach, and 99.81% with SMOTE) except for the under sampling situation, in which the sensitivity was 85.71%. Moreover, the 1D-CNN model incorporating SMOTE yielded promising results in 88% specificity, 93.72% quality index (QI), and 95.10% area under the curve. The model exhibited excellent performance in terms of sensitivity in every scenario except for undersampling. The oversampling of training data with SMOTE yielded a decent level of specificity, demonstrating the model’s strong predictive capacity. In addition, the SMOTE scenario resulted in fewer training epochs, which is another accomplishment.

Clinical Application of Multimodal Sentinel Lymph Node Mapping Method in Patients with Breast Cancer Undergoing Neoadjuvant Chemotherapy: An Interim Analysis.

After neoadjuvant chemotherapy (NAC), the SLN identification rate is lower and has a higher false-negative rate than that at upfront surgery. This clinical trial aimed to confirm the effectiveness of sentinel lymph node (SLN) surgery by determining the lymph node identification rate using multimodal SLN marker methods in patients with advanced breast cancer undergoing NAC. This clinical study is a prospective single-center randomized controlled trial involving patients with breast cancer receiving NAC. Patients are randomized (1:1:1) into arm A that involves the use of radioisotope (RI) plus indocyanine green fluorescence (ICG-F); arm B, RI plus vital dye; and, arm C, ICG-F plus vital dye. A total of 348 patients are needed. An interim analysis was performed on 50% of the patients enrolled. The primary outcome of this trial was the SLN identification rate. Among the 164 total patients (median age 51 years), T2 and N1 were the most common clinical stages. The identification rate of SLN was 95% in arm A, 92% in arm B, and 79% in arm C. To assess superior efficacy, the one-sided endpoint was set at α < 0.0056. Arms A and C showed a difference of 0.1597 in the detection rate (p = 0.0055). The use of ICG-F plus vital dye for SLNB was the least effective. The results show that the choice of tracer should be radioisotope in combination with one of the other tracers to have the highest SLN identification rate when SLNB cannot be implemented conventionally due to the circumstances of each institution.

DNA-AuNP encoders array Integrating a multi-modal coupled analysis approach for noninvasive precision screening of myocardial function injury disease in children

Limited-species serum biomarkers could only partially reflect the information on myocardial function injury disease in children, leading to some important information being inevitably lost, which affects the judgment of the whole information, resulting in a high rate of false positives and negatives. There is a critical unmet need for noninvasive precision screening for myocardial function injury disease in children. Here, we introduced a multi-modal coupled analysis of the DNA-AuNP encoders to achieve this goal. The DNA-AuNP encoders react with the serum specimen from myocardial function injury disease in children to generate the multi-modal response signals, such as the fluorescence spectrum, size distribution, and surface ζ-potential distribution. Interestingly, we used dark-field microscopy to observe the interface changes between the DNA-AuNP encoder and serum at the microscopic level. Different machine learning algorithms analyze the correlation of a clinical outcome with multi-modal response signals from a target disease. As the number of modal response signals increases, our approach monotonically increases screening performance. The three-modal coupled analysis provided an average accuracy of up to 91.7% using 37 serum specimens in children from myocarditis, myocardial injury, and health. This study provides a highly generic tool for noninvasive screening using a drop of serum.

Abstract PO5-23-01: SentiNeo: A feasibility study on a magnetic approach to targeted axillary dissection and sentinel lymph node biopsy after neoadjuvant chemotherapy

Abstract Background Neoadjuvant chemotherapy (NAC) is increasingly used to treat certain breast cancer subtypes, and often, with good axillary response. Axillary management after NAC has become a contentious topic. Low detection rates and high false negative rates for sentinel lymph node biopsies (SLNB) could be due to the chemotherapy-induced fibrosis and blockage of the lymphatic drainage. Complete axillary response raises the dilemma of whether axillary clearance is still necessary after NAC in node positive patients. Targeted axillary dissection (TAD) is a safe alternative but the optimal method for index metastatic lymph node marking and lymphatic mapping of additional SLNs is largely undetermined. In SentiNeo, we study the feasibility and accuracy of using superparamagnetic iron oxides nanoparticles (SPIO) and magnetic seeds for axillary staging after NAC. Methods: Patients from two hospitals undergoing NAC were included: 40 without (cN0) and 40 with (cN+) axillary metastases. For cN0, SPIO was injected before and Technetium (Tc99) after NAC. For cN+, a magnetic clip was inserted into the index metastatic node together with SPIO injection before and Tc99 after NAC. Endpoints were 1) Index node detection and 2) SLN detection and concordance rates. Detection rate was defined as the proportion of patients who had SLNs detected as per tracer (using both magnetometer and gamma probe). Concordance was defined as the number of nodes marked by both SPIO and Tc99, divided by the total number of Tc99 marked nodes. As this is a feasibility study, no calculations for statistical significance were done. Results: All 80 patients have been included and 60 have completed NAC and surgery: 35 cN0 with SLNB and 25 cN+ with TAD. Overall, SLN detection rates (excluding the index nodes) were 83.3% (50/60) for SPIO and 76.7% (46/60) for Tc99. In cN0, SLN detection rate was 85.7% (30/35) for SPIO) and 88.6% (31/35) for Tc99. In cN+, the index node was found in 24 of 25 patients. One patient did not receive a magnetic clip due to administrative error. The SLN detection rate (excluding index nodes) was 80% (20/25) for SPIO) and 60% (15/25) for Tc99. In total, 142 lymph nodes were retrieved, of which 71 contained both tracers. The concordance was 84% (out of 84 nodes detected with Tc99, 71 were also detected with SPIO). However, the reversed concordance was only 55% (out of 129 nodes detected with SPIO, 71 were radioactive). The overall mean (median) number of retrieved nodes were 2.2 (2) for SPIO and 1.4 (1) for Tc99. Conclusions: In this pilot study we have shown that injecting SPIO before NAC is feasible with detection rates comparable to Tc99injected after NAC. More nodes were retrieved with SPIO and furthermore, we hypothesize that these nodes might be more representative than nodes detected with a tracer injected after NAC. Finding the index node using the SentiMag probe was easy. For SLN detection with SPIO the same probe was used, making the magnetic approach to axillary staging an appealing alternative with no need for an extra localisation before surgery or access to nuclear medicine facilities. These findings justify a larger study which is in planning. Table. Citation Format: Kian Chin, Roger Olofsson Bagge, Nushin Mirzaei, Anikó Kovács, Henrik Leonhardt, Pontus Zaar, Andreas Karakatsanis, Eirini Pantiora, Staffan Eriksson, Maria Ekholm, Alastair Thompson, Peter Barry, Michael Boland, Vivian Man, Ava Kwong, Fredrik Wärnberg. SentiNeo: A feasibility study on a magnetic approach to targeted axillary dissection and sentinel lymph node biopsy after neoadjuvant chemotherapy [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO5-23-01.

Probabilistic Prediction of Gastrointestinal Ischemia after Cardiothoracic Surgery.

Gastrointestinal ischemia (GIisch) is challenging to diagnose in patients after cardiothoracic surgery. Computed tomography angiography (CTA) carries substantial false-negative and false-positive rates. The aim of the study was to evaluate if a combination of readily available variables improves the diagnosis of GIisch after cardiothoracic surgery. This retrospective study included patients receiving intensive care after cardiothoracic surgery. GIisch was confirmed by surgical and/or endoscopic findings. A GIisch prediction score was developed using the Spiegelhalter-Knill-Jones system in a training cohort then tested in a validation cohort (patients without obvious signs of GIisch on CTA). The training cohort comprised 125 consecutive patients with suspected GIisch in 2008 to 2019, including 85 with confirmed GIisch. CTA, performed in 92 patients, had a high false-negative rate of 17/60 (28%) and a lower false-positive rate of 7/32 (22%). The score included cardiopulmonary bypass, negatively associated with GIisch, and six variables positively associated with GIisch: intraoperative mean arterial pressure < 50 mm Hg, aspartate aminotransferase > 15 N, lactate increase in 24 hour > 20%, and 3 CTA findings, namely, bowel dilation, bowel wall thickening, and mesenteric vasoconstriction. The area under the receiver operating characteristic was 0.82 (95% confidence interval [CI], 0.51-0.93) in the training cohort and 0.82 (95% CI, 0.68-0.96) in the validation cohort (n = 34 patients). Reliability of the predicted probabilities was greatest for probabilities ≤ 30% or ≥ 70%. In patients receiving intensive care after cardiothoracic surgery, GIisch cannot be ruled out based solely on CTA findings. A scoring system combining CTA findings with other variables may improve the diagnosis of GIisch in this population.

Outcomes of sentinel node biopsy according to MRI response in an association with the subtypes in cN1–3 breast cancer after neoadjuvant systemic therapy, multicenter cohort study

BackgroundThis study investigated the feasibility of sentinel lymph node biopsy (SLNB) after neoadjuvant systemic therapy (NAST) in patients with initially high nodal burden.MethodsIn the multicenter retrospective cohort, 388 individuals with cN1–3 breast cancer who underwent NAST and had SLNB followed by completion axillary lymph node dissection were included. In an external validation cohort, 267 patients with HER2+ or triple-negative breast cancer (TNBC) meeting similar inclusion criteria were included. Primary outcome was the false-negative rates (FNRs) of SLNB according to the MRI response and subtypes. We defined complete MRI responders as patients who experienced disappearance of suspicious features in the breast and axilla after NAST.ResultsIn the multicenter retrospective cohort, 130 (33.5%) of 388 patients were of cN2-3, and 55 (14.2%) of 388 patients showed complete MRI responses. In hormone receptor-positive HER2− (n = 207), complete and non-complete responders had a high FNRs (31.3% [95% CI 8.6–54.0] and 20.9% [95% CI 14.1–27.6], respectively). However, in HER2+ or TNBC (n = 181), the FNR of complete MRI responders was 0% (95% CI 0–0), whereas that of non-complete responders was 33.3% (95% CI 20.8–45.9). When we validated our findings in the external cohort with HER2+ or TNBC (n = 267), of which 34.2% were cN2-3, the FNRs of complete were 7.1% (95% CI 0–16.7).ConclusionsOur findings suggest that SLNB can be a reliable option for nodal status evaluation in selected patients who have responded well to NAST, especially in HER2+ and TNBC patients who show a complete MRI response.

Relevance of the Get Active Questionnaire for Pre-Participation Exercise Screening in the General Population in a Tropical Environment.

The Get Active Questionnaire (GAQ), developed by the Canadian Society for Exercise Professionals (CSEP), was recently recommended for pre-participation screening of the general population in Singapore before increasing their exercise levels. This literature review examines the evidence behind the GAQ and its relevance to our tropical environment. Searches were carried out via Pubmed, MEDLINE and the Cochrane Central Register of Controlled Trials. Resources referenced by the CSEPs were hand searched. The CSEP was also contacted for further information. The evidence behind each GAQ question was compared to international literature and guidelines, where applicable. Out of 273 studies, 49 were suitable for analysis. Two GAQ studies commissioned by the CSEP showed a high negative predictive value but high false negative rate. Of the nine GAQ questions, those on dizziness, joint pains and chronic diseases appear to be justified. Those on heart disease/stroke, hypertension, breathlessness and concussion require modification. The one on syncope can be amalgamated into the dizziness question. The remaining question may be deleted. No long-term studies were available to validate the use of the GAQ. Heat disorders were not considered in the GAQ. Modification of the GAQ, including the inclusion of environmental factors, may make it more suitable for the general population and should be considered.

Risk factors associated with false negative rate of sentinel lymph node biopsy in endometrial cancer: a systematic review and meta-analysis.

Currently, sentinel lymph node biopsy (SLNB) is increasingly used in endometrial cancer, but the rate of missed metastatic lymph nodes compared to systemic lymph node dissection has been a concern. We conducted a systematic review and meta-analysis to evaluate the false negative rate (FNR) of SLNB in patients with endometrial cancer and to explore the risk factors associated with this FNR. Three databases (PubMed, Embase, Web of Science) were searched from initial database build to January 2023 by two independent reviewers. Studies were included if they included 10 or more women diagnosed with International Federation of Gynecology and Obstetrics (FIGO) stage I or higher endometrial cancer, the study technique used sentinel lymph node localization biopsy, and the reported outcome metrics included false negative and/or FNR. Two authors independently reviewed the abstracts and full articles. The FNR and factors associated with FNR were synthesized through random-effects meta-analyses and meta-regression. We identified 62 eligible studies. The overall FNR for the 62 articles was 4% (95% CL 3-5).There was no significant difference in the FNR in patients with high-risk endometrial cancer compared to patients with low-risk endometrial cancer. There was no difference in the FNR for whether frozen sections were used intraoperatively. The type of dye used intraoperatively (indocyanine green/blue dye) were not significantly associated with the false negative rate. Cervical injection reduced the FNR compared with alternative injection techniques. Indocyanine green reduced the FNR compared with alternative Tc-99m. Postoperative pathologic ultrastaging reduced the FNR. Alternative injection techniques (other than the cervix), Tc-99m dye tracer, and the absence of postoperative pathologic ultrastaging are risk factors for a high FNR in endometrial cancer patients who undergo SLNB; therefore, we should be vigilant for missed diagnosis of metastatic lymph nodes after SLNB in such populations. http://www.crd.york.ac.uk/PROSPERO/, identifier CRD42023433637.

Review of methodologies for detecting an observer effect in commercial fisheries data

Observers are deployed on commercial fishing trips to collect representative samples of discard rates. However, fishers may change their fishing habits when an observer is onboard (“observer effect”) or observer programs may over- or under-sample portions of the fleet (“deployment effect”). If the extent of these effects are substantial, observer data will not be representative of unobserved trips, potentially biasing the estimation of discards. This sampling bias can impact catch monitoring, stock assessments, and fishery management. The purpose of this study was to examine the power and error rate of several published methods for detecting an observer effect using a simulation of observer and deployment effects at varying sampling ratios (i.e., observer coverage) for several sample statistics. The simplest methods (t-test and F-test for difference of means and variances) provided an accurate although imprecise estimate of the observer effect size, but only when there were no deployment effects. A generalized linear mixed effects model (GLMM) was also not reliable for detecting small bias, but was not confounded by deployment effects and was relatively robust to changing coverage rates. The most complicated tests involved comparing differences in trip characteristics between subsequent trips for observed-unobserved and unobserved-unobserved pairs. These tests were able to detect smaller observer effects and were not confounded by deployment effects, but were unreliable at high coverage rates (>60%), producing both high false positive and false negative rates. Sensitivity tests also showed differing detection accuracy as the distribution of the metric of interest changed. Thus, the optimal test for detecting an observer effect will depend on the metric of interest, the coverage rate, and whether a deployment effect exists. An example from the New England groundfish fishery is provided to illustrate how conflicting results may be explained. Results should always be considered carefully when declaring that an observer effect is or is not occurring because of the sensitivity of the tests.

Improving the Application of Yolov7 Algorithm in Pin-Losing-Bolts Detection on Power Transmission Lines

A method for bolt deficiency detection on power transmission lines based on YOLOv7 is proposed to address the issues of low detection accuracy and high false negatives in current unmanned aerial vehicle (UAV) inspections. In the Neck module, the existing feature fusion structure is modified by directly combining deep and shallow feature information across adjacent scale feature maps. The prior box sizes in the original YOLOv7 are adjusted to better fit the dimensions and aspect ratios of bolt targets. The improved YOLOv7 algorithm enhances the learning capability of bolt feature information. Experimental results on the bolt deficiency dataset using this method show a 6% increase in precision and a 0.6% increase in mAP_0.5. The experiments demonstrate that the improved method enhances the detection ability of bolt deficiencies on power transmission lines and has practical value in intelligent inspections.

Abstract 4147: Chitosan luminescent rare-earth doped nanoparticles as cancer cell tracking for imaging tumors and their immune response

Abstract Background: Cancer is a group of diseases characterized by the unregulated growth of cells, which can be hard to diagnose due to the high false negatives or positives derived from technical and human mistakes.The detection of solid tumors during resection surgery is also a common issue, so contrast agents are used for both cancer diagnosis and tumor visualization during surgery, facilitating the tumor resection procedure. Nevertheless, they cause long-term problems because they accumulate in specific tissues, low penetration, and rapid clearance from the circulation. It is necessary to study new contrast agents that improve tumor detection to address these issues.Here, we synthesized two different kinds of luminescent nanoparticles (LNPs) conjugated with chitosan (Ch) and functionalized with folic acid (FA) for breast cancer detection. We compared in vitro its physicochemical characteristics and effects on the proliferation of cancer cells. Evaluated its impact on the immune system in vivo and implemented a differential imaging process to resemble a refined 3D visualization of cluster cells using Light Sheet Microscopy. Results: Luminescent nanoparticles (LNPs) synthesized by sol-gel method and functionalized with FA using EDC-NHS coupling revealed high-intensity yellow emission at 540 nm wavelength with a quantum yield of ~ 36%. In vitro viability assay at different times with 293T epithelial cells, MDA-MB-231, and T-47D cancer cells demonstrated biocompatibility and high improvement of cellular uptake by LNP functionalized with FA for cancer cells versus normal cells. Light-sheet microscopy analysis of cell-LNP interactions to visualize tagged MDA-MB-231 clusters with LNP/Ch-FA showed positive interactions due to higher stability and adherence. Biodistribution and effects on the immune system by LNP/Ch-FA nanoparticles inoculated in mice were evaluated by flow cytometer analysis. LNP/Ch-FA were well tolerated in vivo. No adverse effects were observed macroscopically. Peripheral blood analysis from animals injected with LNP/Ch-FA after 2 or 6 days of the treatment showed that a single injection of the different NPs did not have any effects on the amount of circulating neutrophils or monocytes in peripheral blood when assessed 2 or 6 days after the injection. With T cells, the core NP induced a significant decrease in circulating T cells (-31%, p=0.045) that seemed specific to the CD4+ T cells (-33%, p=0.0479). However, conjugation of the NP to chitosan prevented this effect. Conclusions: Our work presented significant evidence that could allow a better understanding of LNPs conjugated with Ch and functionalized with FA for possible upcoming clinical applications that could work singly or in synergy with the present detection techniques to boost accuracy and precision for cancer detection without compromise the inmunosystem. Citation Format: Rodrigo Osorio-Arciniega, Eduardo Lucero-Meza, Bonifacio Can-Uc, Israel Rocha-Mendoza, Pierrick G.J. Fournier, Gustavo A Hirata, Patricia Juarez. Chitosan luminescent rare-earth doped nanoparticles as cancer cell tracking for imaging tumors and their immune response [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4147.

How Accurate Are Anatomical Surface Topography Parameters in Indicating the Presence of a Scoliosis?

Retrospective analysis of a longitudinal cohort. To identify the presence of scoliosis from surface data. Identifying AIS can be difficult. Screening is not universal for reasons including high false positive and negative rates. These difficulties can lead to some adolescents missing out on bracing. Logistic regression analysis of ISIS2 surface topography images only was performed. The x,y positions of the shoulders (Sh), axillae (Ax), waist (waist) and the x,y,z positions of the most prominent points over the posterior torso (scap) were used for the thoracic, thoracolumbar/lumbar and whole spine. The models were used to identify the presence of a 20-degree or larger scoliosis. Differences in the position of the landmarks were analyzed comparing left and right, with the suffix "Ht" representing a difference in the y coordinate, "Off" the x coordinate, and "Depth," the z coordinate. Model accuracy was assessed as both percentages and ROC curves with the coefficients as odds ratios. There were 1283 images (1015 females and 268 males) all with a diagnosis of AIS. The models identified scoliosis in the thoracic spine with an 83% accuracy (AUC 0.91), thoracolumbar/lumbar spine with 74% accuracy (AUC 0.76), and whole spine with 80% accuracy (AUC 0.88). Significant parameters were AxDiffHt, AxDiffOff, WaistDiffHt, ScapDiffOff, and ScapDiffHt for the thoracic curves, AxDiffHt, AxDiffOff, WaistDiffHt for the thoracolumbar/lumbar curves, and AxDiffHt, AxDiffOff, WaistDiffHt and ScapDiffHt for the whole spine. The use of fixed anatomical points around the torso, analyzed using logistic regression, has a high accuracy for identifying curves in the thoracic, thoracolumbar/lumbar, and whole spines. While coming from surface topography images, the results raise the future use of digital photography as a tool for the identification of small scoliosis without using other imaging techniques.

Development of a diagnostic support system for distal humerus fracture using artificial intelligence.

AI has shown promise in automating and improving various tasks, including medical image analysis. Distal humerus fractures are a critical clinical concern that requires early diagnosis and treatment to avoid complications. The standard diagnostic method involves X-ray imaging, but subtle fractures can be missed, leading to delayed or incorrect diagnoses. Deep learning, a subset of artificial intelligence, has demonstrated the ability to automate medical image analysis tasks, potentially improving fracture identification accuracy and reducing the need for additional and cost-intensive imaging modalities (Schwarz et al. 2023). This study aims to develop a deep learning-based diagnostic support system for distal humerus fractures using conventional X-ray images. The primary objective of this study is to determine whether deep learning can provide reliable image-based fracture detection recommendations for distal humerus fractures. Between March 2017 and March 2022, our tertiary hospital's PACS data were evaluated for conventional radiography images of the anteroposterior (AP) and lateral elbow for suspected traumatic distal humerus fractures. The data set consisted of 4931 images of patients sevenyears and older, after excluding paediatric images below sevenyears due to the absence of ossification centres. Two senior orthopaedic surgeons with 12 + years of experience reviewed and labelled the images as fractured or normal. The data set was split into training sets (79.88%) and validation tests (20.1%). Image pre-processing was performed by cropping the images to 224 × 224 pixels around the capitellum, and the deep learning algorithm architecture used was ResNet18. The deep learning model demonstrated an accuracy of 69.14% in the validation test set, with a specificity of 95.89% and a positive predictive value (PPV) of 99.47%. However, the sensitivity was 61.49%, indicating that the model had a relatively high false negative rate. ROC analysis showed an AUC of 0.787 when deep learning AI was the reference and an AUC of 0.580 when the most senior orthopaedic surgeon was the reference. The performance of the model was compared with that of other orthopaedic surgeons of varying experience levels, showing varying levels of diagnostic precision. The developed deep learning-based diagnostic support system shows potential for accurately diagnosing distal humerus fractures using AP and lateral elbow radiographs. The model's specificity and PPV indicate its ability to mark out occult lesions and has a high false positive rate. Further research and validation are necessary to improve the sensitivity and diagnostic accuracy of the model for practical clinical implementation.

Prospective validation of the role of PET/CT in detecting disease after neoadjuvant chemotherapy in advanced ovarian cancer

ObjectivesThe study aimed to compare the diagnostic accuracies of 2-[18F]FDG PET/CT and contrast-enhanced CT (ceCT) after neoadjuvant chemotherapy (NACT) in advanced ovarian cancer (OC).Materials and methodsThis study consisted historical observational cohort and prospective validation cohort. Patients with newly diagnosed stage III–IV OC scheduled for NACT were recruited, with imaging performed after three to six cycles of NACT before interval debulking surgery. Nineteen regions in the abdominopelvic cavity were scored for the presence and absence of disease, referenced to the intra-operative findings or histological specimens. Diagnostic metrics were compared using McNemar’s test.ResultsIn the historical cohort (23 patients, age 58 ± 13), 2-[18F]FDG PET had an overall accuracy (Acc) 82%, sensitivity (Sen) 38%, specificity (Spe) 97%, positive predictive value (PPV) 79% and negative predictive value (NPV) 82%; ceCT had an overall Acc 86%, Sen 64%, Spe 93%, PPV 75% and NPV 89%. In the prospective cohort (46 patients, age 59 ± 9), 2-[18F] FDG PET had an overall Acc 87%, Sen 48%, Spe 98%, PPV 84% and NPV 88%; ceCT had an overall Acc 89%, Sen 66%, Spe 95%, PPV 77% and NPV 91%. No significant difference was demonstrated between the two imaging modalities (p > 0.05). High false-negative rates were observed in the right subdiaphragmatic space, omentum, bowel mesentery and serosa. High omental metabolic uptake after NACT was associated with histological non-responders (p < 0.05).Conclusion2-[18F]FDG PET/CT had no additional value over ceCT with comparable diagnostic accuracy in detecting disease after NACT in advanced OC.Clinical relevance statement2-[18F]FDG PET/CT is not superior to contrast-enhanced CT in determining disease after neoadjuvant chemotherapy in advanced ovarian cancer; contrast-enhanced CT should be suffice for surgical planning before interval debulking surgery.Key Points• Additional value of 2-[18F]FDG PET/CT over contrast-enhanced CT is undefined in detecting disease after neoadjuvant chemotherapy.• 2-[18F]FDG PET/CT has comparable diagnostic accuracy compared to contrast-enhanced CT.• Contrast-enhanced CT will be suffice for surgical planning after neoadjuvant chemotherapy.

Research on neural networks in computer network security evaluation and prediction methods

Anomaly detection in networks to identify intrusions is a common and successful security measure used in many different types of network infrastructure. Network data traffic has increased due to the proliferation of viruses and other forms of cyber-attacks as network technology and applications have developed quickly. The limitations of classical intrusion detection, such as poor detection accuracy, high false negatives, and dependence on dimensionality reduction methods, become more apparent in the face of massive traffic volumes and characteristic information. That’s why IoT infrastructures often use Software-Defined Networking (SDN), allowing for better network adaptability and control. Hence, this paper’s convolutional neural network-based Security Evaluation Model (CNN-SEM) is proposed to secure the source SDN controller from traffic degradation and protect the source network from DDoS assaults. The proposed CNN-SEM system might defend against DDoS assaults once discovered by applying and testing a Convolutional Neural Network (CNN). The model can automatically extract the useful aspects of incursion samples, allowing for precise classification of such data. The detection and mitigation modules evaluate the proposed SDN security system’s performance, and the findings showed promise against next-generation DDoS assaults. The experimental results show the CNN-SEM achieves a high accuracy ratio of 96.6%, a detection ratio of 97.1%, precision ratio of 97.2%, a performance ratio of 95.1% and an enhanced security rate of 98.1% compared to other methods.