10-fold Cross-validation Research Articles

Comparison of genomic prediction methods for early growth traits of Inner Mongolia cashmere goats based on multi trait models.

Inner Mongolia cashmere goat is an excellent livestock breed formed through long-term natural selection and artificial breeding, and is currently a world-class dual-purpose breed producing cashmere and meat. Multi trait animal model is considered to significantly improve the accuracy of genetic evaluation in livestock and poultry, enabling indirect selection between traits. In this study, the pedigree, genotype, environment, and phenotypic records of early growth traits of Inner Mongolia cashmere goats were used to build multi trait animal model., Then three methods including ABLUP, GBLUP, and ssGBLUP wereused to estimate the genetic parameters and genomic breeding values of early growth traits (birth weight, weaning weight, average daily weight gain before weaning, and yearling weight). The accuracy and reliability of genomic estimated breeding value are further evaluated using the five fold cross validation method. The results showed that the heritability of birth weight estimated by three methods was 0.13-0.15, the heritability of weaning weight was 0.13-0.20, heritability of daily weight gain before weaning was 0.11-0.14, and the heritability of yearling weight was 0.09-0.14, all of which belonged to moderate to low heritability. There is a strong positive genetic correlation between weaning weight and daily weight gain before weaning, daily weight gain before weaning and yearling weight, with correlation coefficients of 0.77-0.79 and 0.56-0.67, respectively. The same pattern was found in phenotype correlation among traits. The accuracy of the estimated breeding values by ABLUP, GBLUP, and ssGBLUP methods for birth weight is 0.5047, 0.6694, and 0.7156, respectively; the weaning weight is 0.6207, 0.6456, and 0.7254, respectively; the daily weight gain before weaning was 0.6110, 0.6855, and 0.7357 respectively; and the yearling weight was 0.6209, 0.7155, and 0.7756, respectively. In summary, the early growth traits of Inner Mongolia cashmere goats belong to moderate to low heritability, and the speed of genetic improvement is relatively slow. The genetic improvement of other growth traits can be achieved through the selection of weaning weight. The ssGBLUP method has the highest accuracy and reliability in estimating genomic breeding value of early growth traits in Inner Mongolia cashmere goats, and is significantly higher than that from ABLUP method, indicating that it is the best method for genomic breeding of early growth weight in Inner Mongolia cashmere goats.

Exploring Schizophrenia Classification Through Multimodal MRI and Deep Graph Neural Networks: Unveiling Brain Region-Specific Weight Discrepancies and Their Association With Cell-Type Specific Transcriptomic Features.

Schizophrenia (SZ) is a prevalent mental disorder that imposes significant health burdens. Diagnostic accuracy remains challenging due to clinical subjectivity. To address this issue, we explore magnetic resonance imaging (MRI) as a tool to enhance SZ diagnosis and provide objective references and biomarkers. Using deep learning with graph convolution, we represent MRI data as graphs, aligning with brain structure, and improving feature extraction, and classification. Integration of multiple modalities is expected to enhance classification. Our study enrolled 683 SZ patients and 606 healthy controls from 7 hospitals, collecting structural MRI and functional MRI data. Both data types were represented as graphs, processed by 2 graph attention networks, and fused for classification. Grad-CAM with graph convolution ensured interpretability, and partial least squares analyzed gene expression in brain regions. Our method excelled in the classification task, achieving 83.32% accuracy, 83.41% sensitivity, and 83.20% specificity in 10-fold cross-validation, surpassing traditional methods. And our multimodal approach outperformed unimodal methods. Grad-CAM identified potential brain biomarkers consistent with gene analysis and prior research. Our study demonstrates the effectiveness of deep learning with graph attention networks, surpassing previous SZ diagnostic methods. Multimodal MRI's superiority over unimodal MRI confirms our initial hypothesis. Identifying potential brain biomarkers alongside gene biomarkers holds promise for advancing objective SZ diagnosis and research in SZ.

Radiomics analysis of ultrasound images to discriminate between benign and malignant adnexal masses with solid ultrasound morphology.

Our primary aim was to identify radiomic ultrasound features that can distinguish benign from malignant adnexal masses with solid ultrasound morphology, and primary invasive from metastatic solid ovarian masses, and to develop ultrasound-based machine learning models that include radiomics features to discriminate between benign and malignant solid adnexal masses. Our secondary aim was to compare the diagnostic performance of our radiomics models with that of the ADNEX model and subjective assessment by an experienced ultrasound examiner. This is a retrospective observational single center study. Patients with a histological diagnosis of an adnexal tumor with solid morphology at preoperative ultrasound examination performed between 2014 and 2021 were included. The patient cohort was split into training and validation sets with a ratio of 70:30 and with the same proportion of benign and malignant (borderline, primary invasive and metastatic) tumors in the two subsets. The extracted radiomic features belonged to two different families: intensity-based statistical features and textural features. Models to predict malignancy were built based on a random forest classifier, fine-tuned using 5-fold cross-validation over the training set, and tested on the held-out validation set. The variables used in model building were patient's age, and those radiomic features that were statistically significantly different between benign and malignant adnexal masses (Wilcoxon-Mann-Whitney Test with Benjamini-Hochberg correction for multiple comparisons) and assessed as not redundant based on the Pearson correlation coefficient. We describe discriminative ability as area under the receiver operating characteristics curve (AUC) and classification performance as sensitivity and specificity. 326 patients were identified and 775 preoperative ultrasound images were analyzed. 68 radiomic features were extracted, 52 differed statistically significantly between benign and malignant tumors in the training set, and 18 features were selected for inclusion in model building. The same 52 radiomic features differed statistically significantly between benign, primary invasive malignant and metastatic tumors. However, the values of the features manifested overlap between primary malignant and metastatic tumors and did not differ statistically significantly between them. In the validation set, 25/98 tumors (25.5%) were benign, 73/98 (74.5%) were malignant (6 borderline, 57 primary invasive, 10 metastases). In the validation set, a model including only radiomics features had an AUC of 0.80, and 78% sensitivity and 76% specificity at its optimal risk of malignancy cutoff (68% based on Youden's index). The corresponding results for a model including age and radiomics features were 0.79, 86% and 56% (cutoff 60% based on Youden's method), while those of the ADNEX model were 0.88, 99% and 64% (at 20% malignancy cutoff). Subjective assessment had sensitivity 99% and specificity 72%. Even though our radiomics models had discriminative ability inferior to that of the ADNEX model, our results are promising enough to justify continued development of radiomics analysis of ultrasound images of adnexal masses. This article is protected by copyright. All rights reserved.

FruitVision: A deep learning based automatic fruit grading system

Abstract Quality assessment of fruits plays a key part in the global economy’s agricultural sector. In recent years, it has been shown that fruits are affected by different diseases, which can lead to widespread economic failure in the agricultural industry. Traditional manual visual grading of fruits could be more accurate, making it difficult for agribusinesses to assess quality efficiently. Automatic grading of fruits using computer vision has become a prominent area of study for many researchers. In this study, a deep learning-based model called FruitVision is proposed for the automatic grading of various fruits. The results showed that FruitVision performed all the existing models and obtained an accuracy of 99.42, 99.50, 99.24, 99.12, 99.38, 99.38, 99.17, 98.86, and 97.96% for the apple, banana, guava, lime, orange, pomegranate, Ajwa date, Mabroom date, and mango, respectively, using 5-fold cross-validation. This is a remarkable achievement in the field of AI-based fruit grading systems.

Searching for LINCS to Stress: Using Text Mining to Automate Reference Chemical Curation.

Adaptive stress response pathways (SRPs) restore cellular homeostasis following perturbation but may activate terminal outcomes like apoptosis, autophagy, or cellular senescence if disruption exceeds critical thresholds. Because SRPs hold the key to vital cellular tipping points, they are targeted for therapeutic interventions and assessed as biomarkers of toxicity. Hence, we are developing a public database of chemicals that perturb SRPs to enable new data-driven tools to improve public health. Here, we report on the automated text-mining pipeline we used to build and curate the first version of this database. We started with 100 reference SRP chemicals gathered from published biomarker studies to bootstrap the database. Second, we used information retrieval to find co-occurrences of reference chemicals with SRP terms in PubMed abstracts and determined pairwise mutual information thresholds to filter biologically relevant relationships. Third, we applied these thresholds to find 1206 putative SRP perturbagens within thousands of substances in the Library of Integrated Network-Based Cellular Signatures (LINCS). To assign SRP activity to LINCS chemicals, domain experts had to manually review at least three publications for each of 1206 chemicals out of 181,805 total abstracts. To accomplish this efficiently, we implemented a machine learning approach to predict SRP classifications from texts to prioritize abstracts. In 5-fold cross-validation testing with a corpus derived from the 100 reference chemicals, artificial neural networks performed the best (F1-macro = 0.678) and prioritized 2479/181,805 abstracts for expert review, which resulted in 457 chemicals annotated with SRP activities. An independent analysis of enriched mechanisms of action and chemical use class supported the text-mined chemical associations (p < 0.05): heat shock inducers were linked with HSP90 and DNA damage inducers to topoisomerase inhibition. This database will enable novel applications of LINCS data to evaluate SRP activities and to further develop tools for biomedical information extraction from the literature.

Protein features fusion using attributed network embedding for predicting protein-protein interaction

BackgroundProtein-protein interactions (PPIs) hold significant importance in biology, with precise PPI prediction as a pivotal factor in comprehending cellular processes and facilitating drug design. However, experimental determination of PPIs is laborious, time-consuming, and often constrained by technical limitations.MethodsWe introduce a new node representation method based on initial information fusion, called FFANE, which amalgamates PPI networks and protein sequence data to enhance the precision of PPIs’ prediction. A Gaussian kernel similarity matrix is initially established by leveraging protein structural resemblances. Concurrently, protein sequence similarities are gauged using the Levenshtein distance, enabling the capture of diverse protein attributes. Subsequently, to construct an initial information matrix, these two feature matrices are merged by employing weighted fusion to achieve an organic amalgamation of structural and sequence details. To gain a more profound understanding of the amalgamated features, a Stacked Autoencoder (SAE) is employed for encoding learning, thereby yielding more representative feature representations. Ultimately, classification models are trained to predict PPIs by using the well-learned fusion feature.ResultsWhen employing 5-fold cross-validation experiments on SVM, our proposed method achieved average accuracies of 94.28%, 97.69%, and 84.05% in terms of Saccharomyces cerevisiae, Homo sapiens, and Helicobacter pylori datasets, respectively.ConclusionExperimental findings across various authentic datasets validate the efficacy and superiority of this fusion feature representation approach, underscoring its potential value in bioinformatics.

Faulty cell prediction accuracy comparison of machine learning algorithms using temperature sensor placement optimization approach in immersion cooled Li-ion battery modules

Immersion cooling is a promising thermal management system for LiBs where the cells are submerged in a thermally conductive coolant, thus improving heat dissipation and prolonging battery life. The present study investigates an aligned arranged 4S4P immersion-cooled LiB module to develop a best-fit machine learning (ML) model that could predict the fault position in the module depending on the inputs from the temperature sensors that are optimized under different operating and fault conditions. The Pearson Correlation Coefficient (PCC) feature selection approach is used to optimize the sensors, while the data is generated from numerical simulations on a 4S4P immersion-cooled LiB module. The model validation through internal experimental trials is performed on a 2S2P battery module. The four ML classification algorithms, which include the K Nearest Neighbors, Random Forest (RF), Extreme Gradient, and Long Short Term Memory (LSTM), are trained on the optimized sensors data and are further tested internally and externally to assess their performance on unseen data within and outside the training range. A 5-fold cross-validation process is also implemented, and following a comprehensive comparison of the predictions based on the accuracies and model elapsed time, the best-fit model is identified. The results conclude that while the LSTM model slightly outweighs the other models with an accuracy of 98.18% for the specific external test cases, the RF model is chosen as the best-fit model with a prediction accuracy of 99.7% based on the error metrics and low training time for the internal testing. The outcomes of this present work contribute to the early identification of battery module failures, enhancing safety and reducing costs.

Genomic prediction model optimization for growth traits of olive flounder (Paralichthys olivaceus)

Genomic prediction (GP) has emerged an effective tool for addressing the many shortcomings of traditional selective breeding, thereby enhancing the selection process. In this study, we optimized GP methods using 5-fold cross-validation to estimate genome-estimated breeding values for the weight traits of olive flounder (Paralichthys olivaceus). To accomplish our goal, we determined the parentage of the target broodstock and the ability of 11 prediction models to predict the weight traits of 1.8-year-old olive flounders, which were genotyped using a 70 K single nucleotide polymorphism (SNP) array. Moreover, our optimization efforts toward the predictive ability of genomic best linear unbiased prediction (GBLUP), Bayesian B (BB), and random forest (RF) methods encompassed changes in various aspects such as fixed effects, SNP quantity, population size, and phenotypic data collected at different fish ages. Additionally, we assessed the predictive ability for the total length and body depth of fish using GBLUP, BB, and RF. Among the 11 prediction methods used in this study, the BB (0.675), Elastic Net (0.679), and RF (0.698) methods exhibited the highest predictive abilities, whereas the GBLUP (0.637) method demonstrated the lowest. Incorporating information regarding fish sex as a fixed effect substantially improved the predictive ability of GBLUP and BB. For mean models, utilizing 3000–5000 random SNP markers resulted in a higher predictive ability, similar to that obtained using 50,000 SNPs. Increasing the population size reduced the standard deviation of the predictive ability. Notably, phenotypic records from 1.8-year-old fish exhibited a significantly higher predictive ability than those from the other age groups. Furthermore, GBLUP, BB, and RF provided higher predictive abilities for length (0.655–0.852) and body depth (0.665–0.861). These findings may significantly shape future olive flounder genomic selection programs and offer valuable insights into GP in aquaculture.

Sampling-Based Machine Learning Models for Intrusion Detection in Imbalanced Dataset

Cybersecurity is one of the important considerations when adopting IoT devices in smart applications. Even though a huge volume of data is available, data related to attacks are generally in a significantly smaller proportion. Although machine learning models have been successfully applied for detecting security attacks on smart applications, their performance is affected by the problem of such data imbalance. In this case, the prediction model is preferable to the majority class, while the performance for predicting the minority class is poor. To address such problems, we apply two oversampling techniques and two undersampling techniques to balance the data in different categories. To verify their performance, five machine learning models, namely the decision tree, multi-layer perception, random forest, XGBoost, and CatBoost, are used in the experiments based on the grid search with 10-fold cross-validation for parameter tuning. The results show that both the oversampling and undersampling techniques can improve the performance of the prediction models used. Based on the results, the XGBoost model based on the SMOTE has the best performance in terms of accuracy at 75%, weighted average precision at 82%, weighted average recall at 75%, weighted average F1 score at 78%, and Matthews correlation coefficient at 72%. This indicates that this oversampling technique is effective for multi-attack prediction under a data imbalance scenario.

A Computational Predictor for Accurate Identification of Tumor Homing Peptides by Integrating Sequential and Deep BiLSTM Features.

Cancer remains a severe illness, and current research indicates that tumor homing peptides (THPs) play an important part in cancer therapy. The identification of THPs can provide crucial insights for drug-discovery and pharmaceutical industries as they allow for tailored medication delivery towards cancer cells. These peptides have a high affinity enabling particular receptors present upon tumor surfaces, allowing for the creation of precision medications that reduce off-target consequences and enhance cancer patient treatment results. Wet-lab techniques are considered essential tools for studying THPs; however, they're labor-extensive and time-consuming, therefore making prediction of THPs a challenging task for the researchers. Computational-techniques, on the other hand, are considered significant tools in identifying THPs according to the sequence data. Despite many strategies have been presented to predict new THP, there is still a need to develop a robust method with higher rates of success. In this paper, we developed a novel framework, THP-DF, for accurately identifying THPs on a large-scale. Firstly, the peptide sequences are encoded through various sequential features. Secondly, each feature is passed to BiLSTM and attention layers to extract simplified deep features. Finally, an ensemble-framework is formed via integrating sequential- and deep features which are fed to a support vector machine which with 10-fold cross-validation to carry to validate the efficiency. The experimental results showed that THP-DF worked better on both [Formula: see text] and [Formula: see text] datasets by achieving accuracy of > 95% which are higher than existing predictors both datasets. This indicates that the proposed predictor could be a beneficial tool to precisely and rapidly identify THPs and will contribute to the cutting-edge cancer treatment strategies and pharmaceuticals.

Machine learning in prenatal MRI predicts postnatal ventricular abnormalities in fetuses with isolated ventriculomegaly.

To evaluate the intracranial structures and brain parenchyma radiomics surrounding the occipital horn of the lateral ventricle in normal fetuses (NFs) and fetuses with ventriculomegaly (FVs), as well as to predict postnatally enlarged lateral ventricle alterations in FVs. Between January 2014 and August 2023, 141 NFs and 101 FVs underwent 1.5 T balanced steady-state free precession (BSSFP), including 68 FVs with resolved lateral ventricles (FVM-resolved) and 33 FVs with stable lateral ventricles (FVM-stable). Demographic data and intracranial structures were analyzed. To predict the enlarged ventricle alterations of FVs postnatally, logistic regression models with 5-fold cross-validation were developed based on lateral ventricle morphology, blended-cortical or/and subcortical radiomics characteristics. Validation of the models' performance was conducted using the receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis (DCA). Significant alterations in cerebral structures were observed between NFs and FVs (p < 0.05), excluding the maximum frontal horn diameter (FD). However, there was no notable distinction between the FVM-resolved and FVM-stable groups (all p > 0.05). Based on subcortical-radiomics on the aberrant sides of FVs, this approach exhibited high efficacy in distinguishing NFs from FVs in the training/validation set, yielding an impressive AUC of 1/0.992. With an AUC value of 0.822/0.743 in the training/validation set, the Subcortical-radiomics model demonstrated its ability to predict lateral ventricle alterations in FVs, which had the greatest predictive advantages indicated by DCA. Microstructural alterations in subcortical parenchyma associated with ventriculomegaly can serve as predictive indicators for postnatal lateral ventricle variations in FVs. It is critical to gain pertinent information from a solitary fetal MRI to anticipate postnatal lateral ventricle alterations in fetuses with ventriculomegaly. This approach holds the potential to diminish the necessity for recurrent prenatal ultrasound or MRI examinations. Fetal ventriculomegaly is a dynamic condition that affects postnatal neurodevelopment. Machine learning and subcortical-radiomics can predict postnatal alterations in the lateral ventricle. Machine learning, applied to single-fetal MRI, might reduce required antenatal testing.

SAROS: A dataset for whole-body region and organ segmentation in CT imaging

The Sparsely Annotated Region and Organ Segmentation (SAROS) dataset was created using data from The Cancer Imaging Archive (TCIA) to provide a large open-access CT dataset with high-quality annotations of body landmarks. In-house segmentation models were employed to generate annotation proposals on randomly selected cases from TCIA. The dataset includes 13 semantic body region labels (abdominal/thoracic cavity, bones, brain, breast implant, mediastinum, muscle, parotid/submandibular/thyroid glands, pericardium, spinal cord, subcutaneous tissue) and six body part labels (left/right arm/leg, head, torso). Case selection was based on the DICOM series description, gender, and imaging protocol, resulting in 882 patients (438 female) for a total of 900 CTs. Manual review and correction of proposals were conducted in a continuous quality control cycle. Only every fifth axial slice was annotated, yielding 20150 annotated slices from 28 data collections. For the reproducibility on downstream tasks, five cross-validation folds and a test set were pre-defined. The SAROS dataset serves as an open-access resource for training and evaluating novel segmentation models, covering various scanner vendors and diseases.

BACK-to-MOVE: Machine learning and computer vision model automating clinical classification of non-specific low back pain for personalised management.

Low back pain (LBP) is a major global disability contributor with profound health and socio-economic implications. The predominant form is non-specific LBP (NSLBP), lacking treatable pathology. Active physical interventions tailored to individual needs and capabilities are crucial for its management. However, the intricate nature of NSLBP and complexity of clinical classification systems necessitating extensive clinical training, hinder customised treatment access. Recent advancements in machine learning and computer vision demonstrate promise in characterising NSLBP altered movement patters through wearable sensors and optical motion capture. This study aimed to develop and evaluate a machine learning model (i.e., 'BACK-to-MOVE') for NSLBP classification trained with expert clinical classification, spinal motion data from a standard video alongside patient-reported outcome measures (PROMs). Synchronised video and three-dimensional (3D) motion data was collected during forward spinal flexion from 83 NSLBP patients. Two physiotherapists independently classified them as motor control impairment (MCI) or movement impairment (MI), with conflicts resolved by a third expert. The Convolutional Neural Networks (CNNs) architecture, HigherHRNet, was chosen for effective pose estimation from video data. The model was validated against 3D motion data (subset of 62) and trained on the freely available MS-COCO dataset for feature extraction. The Back-to-Move classifier underwent fine-tuning through feed-forward neural networks using labelled examples from the training dataset. Evaluation utilised 5-fold cross-validation to assess accuracy, specificity, sensitivity, and F1 measure. Pose estimation's Mean Square Error of 0.35 degrees against 3D motion data demonstrated strong criterion validity. Back-to-Move proficiently differentiated MI and MCI classes, yielding 93.98% accuracy, 96.49% sensitivity (MI detection), 88.46% specificity (MCI detection), and an F1 measure of .957. Incorporating PROMs curtailed classifier performance (accuracy: 68.67%, sensitivity: 91.23%, specificity: 18.52%, F1: .800). This study is the first to demonstrate automated clinical classification of NSLBP using computer vision and machine learning with standard video data, achieving accuracy comparable to expert consensus. Automated classification of NSLBP based on altered movement patters video-recorded during routine clinical examination could expedite personalised NSLBP rehabilitation management, circumventing existing healthcare constraints. This advancement holds significant promise for patients and healthcare services alike.

Hourly near-ground NO2 concentration retrieval from geostationary satellite observations

Abstract. Nitrogen dioxide (NO2) is an important contributor to the formation of acid rain, photochemical smog and aerosol particles, which seriously endangers public health. At present, remote sensing of polar-orbiting satellites is a conventional means to obtain large-scale NO2 distribution, but it cannot capture the rapid change because of long revisit periods. The Advanced Himawari Imager (AHI) on the Himawari-8 geostationary satellite has the advantage of high time resolution, which makes it possible to realize near-real-time atmospheric monitoring. Here, based on the absorption characteristics of NO2 in infrared radiation, hourly near-surface NO2 concentrations are retrieved based on the brightness temperature from AHI and auxiliary information such as meteorology and aerosol. The results of 10-fold cross-validation show that NO2 estimations from satellite are in good agreement with in-situ measurements, and their determination coefficient (R2) can reach 0.79. Due to different emission and atmospheric diffusion conditions at different time, the model performance presents a diurnal variation of high accuracy in the noon and afternoon but low accuracy in the morning. Based on the retrieval dataset, it is found that high NO2 concentrations are mainly concentrated in the densely populated and industrial areas such as the North China area. In addition, NO2 pollution mainly occurs in autumn and winter, and the average NO2 concentration in winter is about 1.63 times that in summer in 2021. This study provides a new insight for satellite retrieval of NO2, which is of great significance for real-time pollution monitoring and public health protection.

Machine learning with electrocardiograms to optimize mortality risk stratification in patients with suspected acute coronary syndrome

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): National Heart, Lung, and Blood Institute (NHLBI) and National Center for Advancing Translational Sciences (NCATS). Background The importance of risk stratification in patients with suspected acute coronary syndrome (ACS) extends beyond diagnosis and immediate treatment. It influences the precision of care delivery and the allocation of resources to those at highest risk of adverse events. Purpose We sought to evaluate the prognostic value of electrocardiogram (ECG) feature-based machine learning models to risk-stratify long-term mortality in those with suspected ACS. Methods This was a multicenter prospective observational cohort study of consecutive, non-traumatic patients evaluated at the emergency department for suspected ACS. The derivation cohort included 4,015 patients from a University Medical Center (age 59±16 years, 47% women, 80% training with 10-fold cross-validation and 20% testing). Ascertainment of all-cause death was based on numerous data sources, including the Centers for Disease Control and Prevention’s National Death Index registry. We trained six machine learning methods for survival analysis using 73 morphological ECG features from presenting prehospital 10-second 12-lead ECGs. We used predicted risk scores in a variational Bayesian Gaussian mixture model to define three risk groups corresponding to low, moderate, and high-risk and compared the resulting classification performance against the HEART score. The external testing cohort included 3,095 patients from a University (age 59±15 years, 44% women). Results The mortality rate was 20.3% in the derivation cohort after an average follow-up period of 3.53 years (IQR 1.75 - 5.32). Extra Survival Trees outperformed other forecasting models during cross-validation. In the internal testing cohort, the derived risk groups were significantly predictive of survival (Kaplan-Meier log-rank test statistic = 121.14, p&amp;lt;0.001), outperforming the classification performance using the HEART score (Figure 1). Our machine learning-based risk stratification model detected &amp;gt;90% of death events missed by low-risk HEART score (Figure 2). Ruling-out the low-risk group, it achieved a negative predictive value of 93.4% with a sensitivity of 85.9% (compared to 89.0% and 75.0%, respectively, for the HEART score). This classification performance generalized well to our external testing cohort. Compared to those in the low-risk group, patients at moderate (OR = 3.62 [1.35-9.74]) and high (OR = 6.12 [2.38-15.75]) risk had significantly higher odds of 30-day cardiovascular death (n=109, 3.5%). Conclusions Using only features from the 10-second 12-lead ECG, we derived and externally validated a machine learning model that stratifies the mortality risk of patients during long-term follow up. This model outperformed current standard of care based on the HEART score. Future work should validate the potential of this decision-support tool in influencing treatment plans and resource allocation.Figure 1.Clustering performanceFigure 2.Sankey diagram

An improved attention module based on nnU-Net for segmenting primary central nervous system lymphoma (PCNSL) in MRI images1.

Accurate volumetric segmentation of primary central nervous system lymphoma (PCNSL) is essential for assessing and monitoring the tumor before radiotherapy and the treatment planning. The tedious manual segmentation leads to interindividual and intraindividual differences, while existing automatic segmentation methods cause under-segmentation of PCNSL due to the complex and multifaceted nature of the tumor. To address the challenges of small size, diffused distribution, poor inter-layer continuity on the same axis, and tendency for over-segmentation in brain MRI PCNSL segmentation, we propose an improved attention module based on nnUNet for automated segmentation. We collected 114 T1 MRI images of patients in the Huashan Hospital, Shanghai. Then randomly split the total of 114 cases into 5 distinct training and test sets for a 5-fold cross-validation. To efficiently and accurately delineate the PCNSL, we proposed an improved attention module based on nnU-Net with 3D convolutions, batch normalization, and residual attention (res-attention) to learn the tumor region information. Additionally, multi-scale dilated convolution kernels with different dilation rates were integrated to broaden the receptive field. We further used attentional feature fusion with 3D convolutions (AFF3D) to fuse the feature maps generated by multi-scale dilated convolution kernels to reduce under-segmentation. Compared to existing methods, our attention module improves the ability to distinguish diffuse and edge enhanced types of tumors; and the broadened receptive field captures tumor features of various scales and shapes more effectively, achieving a 0.9349 Dice Similarity Coefficient (DSC). Quantitative results demonstrate the effectiveness of the proposed method in segmenting the PCNSL. To our knowledge, this is the first study to introduce attention modules into deep learning for segmenting PCNSL based on brain magnetic resonance imaging (MRI), promoting the localization of PCNSL before radiotherapy.

Monitoring saltwater intrusion to estuaries based on UAV and satellite imagery with machine learning models

Saltwater intrusion is a natural mixture process between watershed freshwater and seawater that frequently occurs in estuaries. Station-based monitoring of saltwater intrusion is time-consuming and labor-intensive. To enable quick monitoring of saltwater intrusion, this study developed new remote sensing algorithms for water surface salinity measurement using four decision tree-based machine learning models. These models were built based on simultaneously collected in-situ salinity data from a waterway in the Pearl River Delta and Unmanned Aerial Vehicle (UAV) hyperspectral images. A 10-fold cross-validation was applied to assess the performance of the models, with XGBoost outperforming the other three models (R2=0.93, RMSE = 0.88 psu). Then the developed model was employed for Sentinel-2 multispectral satellite images to invert the estuarine salinity distribution at a larger spatial scale. Results displayed the high performance of the machine learning models proposed in this study for mapping the salinity distribution in river channels, making it an efficient and practical technique for monitoring saltwater intrusion in river channels at a regional scale.

Liver Observation Segmentation on Contrast-Enhanced MRI: SAM and MedSAM Performance in Patients With Probable or Definite Hepatocellular Carcinoma.

Purpose: To evaluate factors impacting the Segment Anything Model (SAM) and variant MedSAM performance for segmenting liver observations on contrast-enhanced (CE) magnetic resonance imaging (MRI) in high-risk patients with probable hepatocellular carcinoma (HCC) (LR-4) and definite HCC (LR-5). Methods: A retrospective cohort of liver observations (LR-4/LR-5) on CE-MRI from 97 patients at high-risk for HCC was derived (2013-2018). Using bounding-boxes as prompts under 5-fold cross-validation, segmentation performance was evaluated at the model and liver observation-levels for: (1) model types: SAM versus MedSAM, (2) image sizes: 256 × 256 versus 512 × 512, (3) image channel composition: CE sequences at 3 phases of enhancement independently and combined, (4) liver observation size: >10 mm versus >20 mm, (5) certainty of diagnosis: LR-4 versus LR-5, and (6) contrast-agent type: hepatobiliary versus extracellular. Segmentation performance, quantified using Dice coefficient, were compared using univariate (Wilcoxon signed-rank and t-test) and multivariable analyses (multiple correspondence analysis and subsequent linear modelling). Results: MedSAM trained on 512 × 512 combined CE sequences performed best with mean Dice coefficient 0.68 (95% confidence interval 0.66, 0.69). Overall, all factors except contrast-agent type affected performance, with larger image size resulting in the highest performance improvement (512 × 512: 0.57, 256 × 256: 0.26, P < .001) at the model-level. Contrast-agents affected performance for patients with LR-4 observations using MedSAM-based models (P < .03). Larger observation size, image size, and higher certainty of diagnosis were associated with better segmentation on multivariable analysis. Conclusion: A variety of factors were found to impact SAM/MedSAM performance for segmenting liver observations in patients with probable and definite HCC on CE-MRI. Future models may be optimized by accounting for these factors.

Predicting Outcomes Following Lower Extremity Endovascular Revascularization Using Machine Learning.

Lower extremity endovascular revascularization for peripheral artery disease carries nonnegligible perioperative risks; however, outcome prediction tools remain limited. Using machine learning, we developed automated algorithms that predict 30-day outcomes following lower extremity endovascular revascularization. The National Surgical Quality Improvement Program targeted vascular database was used to identify patients who underwent lower extremity endovascular revascularization (angioplasty, stent, or atherectomy) for peripheral artery disease between 2011 and 2021. Input features included 38 preoperative demographic/clinical variables. The primary outcome was 30-day postprocedural major adverse limb event (composite of major reintervention, untreated loss of patency, or major amputation) or death. Data were split into training (70%) and test (30%) sets. Using 10-fold cross-validation, 6 machine learning models were trained using preoperative features. The primary model evaluation metric was area under the receiver operating characteristic curve. Overall, 21 886 patients were included, and 30-day major adverse limb event/death occurred in 1964 (9.0%) individuals. The best performing model for predicting 30-day major adverse limb event/death was extreme gradient boosting, achieving an area under the receiver operating characteristic curve of 0.93 (95% CI, 0.92-0.94). In comparison, logistic regression had an area under the receiver operating characteristic curve of 0.72 (95% CI, 0.70-0.74). The calibration plot showed good agreement between predicted and observed event probabilities with a Brier score of 0.09. The top 3 predictive features in our algorithm were (1) chronic limb-threatening ischemia, (2) tibial intervention, and (3) congestive heart failure. Our machine learning models accurately predict 30-day outcomes following lower extremity endovascular revascularization using preoperative data with good discrimination and calibration. Prospective validation is warranted to assess for generalizability and external validity.

Radiomics analysis for prediction and classification of submucosal tumors based on gastrointestinal endoscopic ultrasonography.

To identify and classify submucosal tumors by building and validating a radiomics model with gastrointestinal endoscopic ultrasonography (EUS) images. A total of 144 patients diagnosed with submucosal tumors through gastrointestinal EUS were collected between January 2019 and October 2020. There are 1952 radiomic features extracted from each patient's EUS images. The statistical test and the customized least absolute shrinkage and selection operator regression were used for feature selection. Subsequently, an extremely randomized trees algorithm was utilized to construct a robust radiomics classification model specifically tailored for gastrointestinal EUS images. The performance of the model was measured by evaluating the area under the receiver operating characteristic curve. The radiomics model comprised 30 selected features that showed good discrimination performance in the validation cohorts. During validation, the area under the receiver operating characteristic curve was calculated as 0.9203 and the mean value after 10-fold cross-validation was 0.9260, indicating excellent stability and calibration. These results confirm the clinical utility of the model. Utilizing the dataset provided curated from gastrointestinal EUS examinations at our collaborating hospital, we have developed a well-performing radiomics model. It can be used for personalized and non-invasive prediction of the type of submucosal tumors, providing physicians with aid for early treatment and management of tumor progression.