Model Development Dataset Research Articles

Simulated outcomes for durotomy repair in minimally invasive spine surgery

Minimally invasive spine surgery (MISS) is increasingly performed using endoscopic and microscopic visualization, and the captured video can be used for surgical education and development of predictive artificial intelligence (AI) models. Video datasets depicting adverse event management are also valuable, as predictive models not exposed to adverse events may exhibit poor performance when these occur. Given that no dedicated spine surgery video datasets for AI model development are publicly available, we introduce Simulated Outcomes for Durotomy Repair in Minimally Invasive Spine Surgery (SOSpine). A validated MISS cadaveric dural repair simulator was used to educate neurosurgery residents, and surgical microscope video recordings were paired with outcome data. Objects including durotomy, needle, grasper, needle driver, and nerve hook were then annotated. Altogether, SOSpine contains 15,698 frames with 53,238 annotations and associated durotomy repair outcomes. For validation, an AI model was fine-tuned on SOSpine video and detected surgical instruments with a mean average precision of 0.77. In summary, SOSpine depicts spine surgeons managing a common complication, providing opportunities to develop surgical AI models.

The PolitiFact-Oslo Corpus: A New Dataset for Fake News Analysis and Detection

This study presents a new dataset for fake news analysis and detection, namely, the PolitiFact-Oslo Corpus. The corpus contains samples of both fake and real news in English, collected from the fact-checking website PolitiFact.com. It grew out of a need for a more controlled and effective dataset for fake news analysis and detection model development based on recent events. Three features make it uniquely placed for this: (i) the texts have been individually labelled for veracity by experts, (ii) they are complete texts that strictly correspond to the claims in question, and (iii) they are accompanied by important metadata such as text type (e.g., social media, news and blog). In relation to this, we present a pipeline for collecting quality data from major fact-checking websites, a procedure which can be replicated in future corpus building efforts. An exploratory analysis based on sentiment and part-of-speech information reveals interesting differences between fake and real news as well as between text types, thus highlighting the importance of adding contextual information to fake news corpora. Since the main application of the PolitiFact-Oslo Corpus is in automatic fake news detection, we critically examine the applicability of the corpus and another PolitiFact dataset built based on less strict criteria for various deep learning-based efficient approaches, such as Bidirectional Long Short-Term Memory (Bi-LSTM), LSTM fine-tuned transformers such as Bidirectional Encoder Representations from Transformers (BERT) and RoBERTa, and XLNet.

Automatic Segmentation and Quantification of Abdominal Aortic Calcification in Lateral Lumbar Radiographs Based on Deep-Learning-Based Algorithms.

To investigate the performance of deep-learning-based algorithms for the automatic segmentation and quantification of abdominal aortic calcification (AAC) in lateral lumbar radiographs, we retrospectively collected 1359 consecutive lateral lumbar radiographs. The data were randomly divided into model development and hold-out test datasets. The model development dataset was used to develop U-shaped fully convolutional network (U-Net) models to segment the landmarks of vertebrae T12-L5, the aorta, and anterior and posterior aortic calcifications. The AAC lengths were calculated, resulting in an automatic Kauppila score output. The vertebral levels, AAC scores, and AAC severity were obtained from clinical reports and analyzed by an experienced expert (reference standard) and the model. Compared with the reference standard, the U-Net model demonstrated a good performance in predicting the total AAC score in the hold-out test dataset, with a correlation coefficient of 0.97 (p <0.001). The overall accuracy for the AAC severity was 0.77 for the model and 0.74 for the clinical report. Additionally, the Kendall coefficient of concordance of the total AAC score prediction was 0.89 between the model-predicted score and the reference standard, and 0.88 between the structured clinical report and the reference standard. In conclusion, the U-Net-based deep learning approach demonstrated a relatively high model performance in automatically segmenting and quantifying ACC.

Developing an Artificial Intelligence-Based Pediatric and Adolescent Migraine Diagnostic Model.

Introduction Misdiagnosis of pediatric and adolescent migraine is a significant problem. The first artificial intelligence (AI)-based pediatric migraine diagnosis model was made utilizing a database of questionnaires obtained from a previous epidemiological study, the Itoigawa Benizuwaigani Study. Methods The AI-based headache diagnosis model was created based on the internal validation based on a retrospective investigation of 909 patients (636 training dataset for model development and 273 test dataset for internal validation) aged six to 17 years diagnosed based on the International Classification of Headache Disorders 3rd edition. The diagnostic performance of the AI model was evaluated. Results The dataset included 234/909 (25.7%) pediatric or adolescent patients with migraine. The mean age was 11.3 (standard deviation 3.17) years. The model's accuracy, sensitivity (recall), specificity, precision, and F-values for the test dataset were 94.5%, 88.7%, 96.5%, 90.0%, and 89.4%, respectively. Conclusions The AI model exhibited high diagnostic performance for pediatric and adolescent migraine. It holds great potential as a powerful tool for diagnosing these conditions, especially when secondary headaches are ruled out. Nonetheless, further data collection and external validation are necessary to enhance the model's performance and ensure its applicability in real-world settings.

Developing an artificial intelligence-based headache diagnostic model from a clinic patients' dataset.

We developed an artificial intelligence (AI)-based headache diagnosis model using a large questionnaire database in a headache-specializing clinic. Misdiagnosis of headache disorders is a serious issue and AI-based headache diagnosis models are scarce. We developed an AI-based headache diagnosis model and conducted internal validation based on a retrospective investigation of 6058 patients (4240 training dataset for model development and 1818 test dataset for internal validation) diagnosed by a headache specialist. The ground truth was the diagnosis by the headache specialist. The diagnostic performance of the AI model was evaluated. The dataset included 4829/6058 (79.7%) patients with migraine, 834/6058 (13.8%) with tension-type headache, 78/6058 (1.3%) with trigeminal autonomic cephalalgias, 38/6058 (0.6%) with other primary headache disorders, and 279/6058 (4.6%) with other headaches. The mean (standard deviation) age was 34.7 (14.5) years, and 3986/6058 (65.8%) were female. The model's micro-average accuracy, sensitivity (recall), specificity, precision, and F-values for the test dataset were 93.7%, 84.2%, 84.2%, 96.1%, and 84.2%, respectively. The diagnostic performance for migraine was high, with a sensitivity of 88.8% and c-statistics of 0.89 (95% confidence interval 0.87-0.91). Our AI model demonstrated high diagnostic performance for migraine. If secondary headaches can be ruled out, the model can be a powerful tool for diagnosing migraine; however, further data collection and external validation are required to strengthen the performance, ensure the generalizability in other outpatients, and demonstrate its utility in real-world settings.

Development and Validation of an Automated Image-Based Deep Learning Platform for Sarcopenia Assessment in Head and Neck Cancer

Sarcopenia is an established prognostic factor in patients with head and neck squamous cell carcinoma (HNSCC); the quantification of sarcopenia assessed by imaging is typically achieved through the skeletal muscle index (SMI), which can be derived from cervical skeletal muscle segmentation and cross-sectional area. However, manual muscle segmentation is labor intensive, prone to interobserver variability, and impractical for large-scale clinical use. To develop and externally validate a fully automated image-based deep learning platform for cervical vertebral muscle segmentation and SMI calculation and evaluate associations with survival and treatment toxicity outcomes. For this prognostic study, a model development data set was curated from publicly available and deidentified data from patients with HNSCC treated at MD Anderson Cancer Center between January 1, 2003, and December 31, 2013. A total of 899 patients undergoing primary radiation for HNSCC with abdominal computed tomography scans and complete clinical information were selected. An external validation data set was retrospectively collected from patients undergoing primary radiation therapy between January 1, 1996, and December 31, 2013, at Brigham and Women's Hospital. The data analysis was performed between May 1, 2022, and March 31, 2023. C3 vertebral skeletal muscle segmentation during radiation therapy for HNSCC. Overall survival and treatment toxicity outcomes of HNSCC. The total patient cohort comprised 899 patients with HNSCC (median [range] age, 58 [24-90] years; 140 female [15.6%] and 755 male [84.0%]). Dice similarity coefficients for the validation set (n = 96) and internal test set (n = 48) were 0.90 (95% CI, 0.90-0.91) and 0.90 (95% CI, 0.89-0.91), respectively, with a mean 96.2% acceptable rate between 2 reviewers on external clinical testing (n = 377). Estimated cross-sectional area and SMI values were associated with manually annotated values (Pearson r = 0.99; P < .001) across data sets. On multivariable Cox proportional hazards regression, SMI-derived sarcopenia was associated with worse overall survival (hazard ratio, 2.05; 95% CI, 1.04-4.04; P = .04) and longer feeding tube duration (median [range], 162 [6-1477] vs 134 [15-1255] days; hazard ratio, 0.66; 95% CI, 0.48-0.89; P = .006) than no sarcopenia. This prognostic study's findings show external validation of a fully automated deep learning pipeline to accurately measure sarcopenia in HNSCC and an association with important disease outcomes. The pipeline could enable the integration of sarcopenia assessment into clinical decision making for individuals with HNSCC.

The Evolution and Reliability of Machine Learning Techniques for Oncology

It is no secret that the rise of the Internet and other digital technologies has sparked renewed interest in AI-based techniques, especially those that fall under the umbrella of the subset of algorithms known as "Machine Learning" (ML).&#x0D; These advancements in electronics have allowed us to comprehend the world beyond the bounds of human cognition. A high-dimensional dataset's complicated nature. Although these techniques have been regularly employed by the medical sciences, their adoption to enhance patient care has been a bit slow. The availability of curated diverse data sets for model development is all examples of the substantial hurdles that have delayed these efforts. The future clinical acceptance of each of these characteristics may be affected by a number of limiting conditions, such as the time and resources spent on data collection and model development, the cost of integration relative to the time and resources spent on translation, and the potential for patient damage. In order to preserve value and enhance medical care, the goal of this article is to evaluate all facets of the issue in light of the validity of using ML methods in cancer, to serve as a template for further research and the subfield of oncology that serves as a model for other parts of the discipline.

Deep learning for image classification between primary central nervous system lymphoma and glioblastoma in corpus callosal tumors.

It can be challenging in some situations to distinguish primary central nervous system lymphoma (PCNSL) from glioblastoma (GBM) based on magnetic resonance imaging (MRI) scans, especially those involving the corpus callosum. The objective of this study was to assess the diagnostic performance of deep learning (DL) models between PCNSLs and GBMs in corpus callosal tumors. The axial T1-weighted gadolinium-enhanced MRI scans of 274 individuals with pathologically confirmed PCNSL (n = 94) and GBM (n = 180) were examined. After image pooling, pre-operative MRI scans were randomly split with an 80/20 procedure into a training dataset (n = 709) and a testing dataset (n = 177) for DL model development. Therefore, the DL model was deployed as a web application and validated with the unseen images (n = 114) and area under the receiver operating characteristic curve (AUC); other outcomes were calculated to assess the discrimination performance. The first baseline DL model had an AUC of 0.77 for PCNSL when evaluated with unseen images. The 2nd model with ridge regression regularization and the 3rd model with drop-out regularization increased an AUC of 0.83 and 0.84. In addition, the last model with data augmentation yielded an AUC of 0.57. DL with regularization may provide useful diagnostic information to help doctors distinguish PCNSL from GBM.

The raw sea surface salinity dataset for RF model development

The raw sea surface salinity dataset for RF model development

Predicting streamflow with LSTM networks using global datasets

Streamflow predictions remain a challenge for poorly gauged and ungauged catchments. Recent research has shown that deep learning methods based on Long Short-Term Memory (LSTM) cells outperform process-based hydrological models for rainfall-runoff modeling, opening new possibilities for prediction in ungauged basins (PUB). These studies usually feature local datasets for model development, while predictions in ungauged basins at a global scale require training on global datasets. In this study, we develop LSTM models for over 500 catchments from the CAMELS-US data base using global ERA5 meteorological forcing and global catchment characteristics retrieved with the HydroMT tool. Comparison against an LSTM trained with local datasets shows that, while the latter generally yields superior performances due to the higher spatial resolution meteorological forcing (overall median daily NSE 0.54 vs. 0.71), training with ERA5 results in higher NSE in most catchments of Western and North-Western US (median daily NSE of 0.83 vs. 0.78). No significant changes in performance occur when substituting local with global data sources for deriving the catchment characteristics. These results encourage further research to develop LSTM models for worldwide predictions of streamflow in ungauged basins using available global datasets. Promising directions include training the models with streamflow data from different regions of the world and with higher quality meteorological forcing.

The feasibility of early detecting coronary artery disease using deep learning-based algorithm based on electrocardiography.

Coronary Artery Disease (CAD) is a major cause of morbidity and mortality, yet it is frequently asymptomatic in the early stages and hence goes undetected. We aimed to develop a novel artificial intelligence-based approach for early detection of CAD patients based solely on electrocardiogram (ECG). This study included patients with suspected CAD who had standard 10-s resting 12-lead ECGs and coronary computed tomography angiography (cCTA) results within 4 weeks or less. The ECG and cCTA data from the same patient were matched based on their hospitalization or outpatient ID. All matched data pairs were then randomly divided into training, validation dataset for model development based on convolutional neural network (CNN) and test dataset for model evaluation. The accuracy (Acc), specificity (Spec), sensitivity (Sen), positive predictive value (PPV), negative predictive value (NPV) and area under the receiver operating characteristic curve (AUC) of the model were calculated by using the test dataset. In the test dataset, the model for detecting CAD achieved an AUC of 0.75 (95% CI, 0.73 to 0.78) with an accuracy of 70.0%. Using the optimal cut-off point, the CAD detection model had sensitivity of 68.7%, specificity of 70.9%, positive predictive value (PPV) of 61.2%, and negative predictive value (NPV) of 77.2%. Our study demonstrates that a well-trained CNN model based solely on ECG could be considered an efficient, low-cost, and noninvasive method of assisting in CAD detection.

Construction and Validation of a Model for Predicting Impaired Fasting Glucose Based on More Than 4000 General Population.

Impaired fasting glucose (IFG) is associated with an increased risk of multiple diseases. Therefore, the early identification and intervention of IFG are particularly significant. Our study aims to construct and validate a clinical and laboratory-based nomogram (CLN) model for predicting IFG risk. This cross-sectional study collected information on health check-up subjects. Risk predictors were screened mainly by the LASSO regression analysis and were applied to construct the CLN model. Furthermore, we showed examples of applications. Then, the accuracy of the CLN model was evaluated by the receiver operating characteristic (ROC) curve, the area under the ROC curve (AUC) values, and the calibration curve of the CLN model in the training set and validation set, respectively. The decision curve analysis (DCA) was used to estimate the level of clinical benefit. Furthermore, the performance of the CLN model was evaluated in the independent validation dataset. In the model development dataset, 2340 subjects were randomly assigned to the training set (N = 1638) and validation set (N = 702). Six predictors significantly associated with IFG were screened and used in the construction of the CLN model, a subject was randomly selected, and the risk of developing IFG was predicted to be 83.6% by using the CLN model. The AUC values of the CLN model were 0.783 in the training set and 0.789 in the validation set. The calibration curve demonstrated good concordance. DCA showed that the CLN model has good clinical application. We further performed independent validation (N = 1875), showed an AUC of 0.801, with the good agreement and clinical diagnostic value. We developed and validated the CLN model that could predict the risk of IFG in the general population. It not only facilitates the diagnosis and treatment of IFG but also helps to reduce the medical and economic burdens of IFG-related diseases.

Machine learning for detecting DNA attachment on SPR biosensor

Optoelectric biosensors measure the conformational changes of biomolecules and their molecular interactions, allowing researchers to use them in different biomedical diagnostics and analysis activities. Among different biosensors, surface plasmon resonance (SPR)-based biosensors utilize label-free and gold-based plasmonic principles with high precision and accuracy, allowing these gold-based biosensors as one of the preferred methods. The dataset generated from these biosensors are being used in different machine learning (ML) models for disease diagnosis and prognosis, but there is a scarcity of models to develop or assess the accuracy of SPR-based biosensors and ensure a reliable dataset for downstream model development. Current study proposed innovative ML-based DNA detection and classification models from the reflective light angles on different gold surfaces of biosensors and associated properties. We have conducted several statistical analyses and different visualization techniques to evaluate the SPR-based dataset and applied t-SNE feature extraction and min-max normalization to differentiate classifiers of low-variances. We experimented with several ML classifiers, namely support vector machine (SVM), decision tree (DT), multi-layer perceptron (MLP), k-nearest neighbors (KNN), logistic regression (LR) and random forest (RF) and evaluated our findings in terms of different evaluation metrics. Our analysis showed the best accuracy of 0.94 by RF, DT and KNN for DNA classification and 0.96 by RF and KNN for DNA detection tasks. Considering area under the receiver operating characteristic curve (AUC) (0.97), precision (0.96) and F1-score (0.97), we found RF performed best for both tasks. Our research shows the potentiality of ML models in the field of biosensor development, which can be expanded to develop novel disease diagnosis and prognosis tools in the future.

Negative Samples for Improving Object Detection-A Case Study in AI-Assisted Colonoscopy for Polyp Detection.

Deep learning object-detection models are being successfully applied to develop computer-aided diagnosis systems for aiding polyp detection during colonoscopies. Here, we evidence the need to include negative samples for both (i) reducing false positives during the polyp-finding phase, by including images with artifacts that may confuse the detection models (e.g., medical instruments, water jets, feces, blood, excessive proximity of the camera to the colon wall, blurred images, etc.) that are usually not included in model development datasets, and (ii) correctly estimating a more realistic performance of the models. By retraining our previously developed YOLOv3-based detection model with a dataset that includes 15% of additional not-polyp images with a variety of artifacts, we were able to generally improve its F1 performance in our internal test datasets (from an average F1 of 0.869 to 0.893), which now include such type of images, as well as in four public datasets that include not-polyp images (from an average F1 of 0.695 to 0.722).

Deep-Learning Models for Detection and Localization of Visible Clinically Significant Prostate Cancer on Multi-Parametric MRI.

Deep learning for diagnosing clinically significant prostate cancer (csPCa) is feasible but needs further evaluation in patients with prostate-specific antigen (PSA) levels of 4-10 ng/mL. To explore diffusion-weighted imaging (DWI), alone and in combination with T2-weighted imaging (T2WI), for deep-learning-based models to detect and localize visible csPCa. Retrospective. One thousand six hundred twenty-eight patients with systematic and cognitive-targeted biopsy-confirmation (1007 csPCa, 621 non-csPCa) were divided into model development (N=1428) and hold-out test (N=200) datasets. DWI with diffusion-weighted single-shot gradient echo planar imaging sequence and T2WI with T2-weighted fast spin echo sequence at 3.0-T and 1.5-T. The ground truth of csPCa was annotated by two radiologists in consensus. A diffusion model, DWI and apparent diffusion coefficient (ADC) as input, and a biparametric model (DWI, ADC, and T2WI as input) were trained based on U-Net. Three radiologists provided the PI-RADS (version 2.1) assessment. The performances were determined at the lesion, location, and the patient level. The performance was evaluated using the areas under the ROC curves (AUCs), sensitivity, specificity, and accuracy. A P value <0.05 was considered statistically significant. The lesion-level sensitivities of the diffusion model, the biparametric model, and the PI-RADS assessment were 89.0%, 85.3%, and 90.8% (P=0.289-0.754). At the patient level, the diffusion model had significantly higher sensitivity than the biparametric model (96.0% vs. 90.0%), while there was no significant difference in specificity (77.0%. vs. 85.0%, P=0.096). For location analysis, there were no significant differences in AUCs between the models (sextant-level, 0.895 vs. 0.893, P=0.777; zone-level, 0.931 vs. 0.917, P=0.282), and both models had significantly higher AUCs than the PI-RADS assessment (sextant-level, 0.734; zone-level, 0.863). The diffusion model achieved the best performance in detecting and localizing csPCa in patients with PSA levels of 4-10 ng/mL. 3 TECHNICAL EFFICACY: Stage 2.

Deep learning–assisted colonoscopy images for prediction of mismatch repair deficiency in colorectal cancer.

24 Background: Deficient mismatch repair (or microsatellite instability) is a major predictive biomarker for the efficacy of immune checkpoint inhibitors of colorectal cancer. However, routine testing is time-consuming and incur substantial costs. We developed and validated a deep learning-based classifiers (MMR-Scopy) to detect mismatch repair-deficient status from routine colonoscopy images. Methods: We retrospectively obtained the colonoscopy images from the imaging database at Endoscopic Center of the Sixth Affiliated Hospital of Sun Yat-sen University. Colorectal cancer images from patients that had undergone colonoscopy procedures, surgery and the mismatch repair immunohistochemistry test were eligible for this study. Colonoscopy images from a prospective trial (Neoadjuvant PD-1 blockade by toripalimab with or without celecoxib in mismatch repair-deficient or microsatellite instability-high locally advanced colorectal cancer, PICC) were used to test MMR-Scopy. The primary outcome was dMMR status. The primary performance metrics were accuracy, negative predictive value (NPV), and area under the receiver operating characteristic curve (AUROC). Results: A total of 5226 eligible images from 892 tumors from the consecutive patients were utilized to develop and validate the deep learning model. 2105 CRC images from 306 tumors were randomly selected to form model development dataset with a class-balanced approach. 3121 images of 488 pMMR tumors and 98 dMMR tumors were used to form the independent dataset. The MMR-Scopy model achieved an AUROC of 0.948 (95% CI 0.919- 0.977) on the test dataset. On the independent validation dataset, the AUROC was 0.807 (0.760- 0.854), and the NPV in was 94.2% (95% CI 0.918-0.967). On the PICC dataset, the model identified 29 tumors among the 33 dMMR tumors (87.88%). Conclusions: MMR-Scopy achieved a high NPV in detecting dMMR colorectal cancers. This model might serve as an automatic screening tool, one that might potentially assist doctors in determining a new treatment strategy. [Table: see text]

Distinguishing common renal cell carcinomas from benign renal tumors based on machine learning: comparing various CT imaging phases, slices, tumor sizes, and ROI segmentation strategies.

To determine whether a CT-based machine learning (ML) can differentiate benign renal tumors from renal cell carcinomas (RCCs) and improve radiologists' diagnostic performance, and evaluate the impact of variable CT imaging phases, slices, tumor sizes, and region of interest (ROI) segmentation strategies. Patients with pathologically proven RCCs and benign renal tumors from our institution between 2008 and 2020 were included as the training dataset for ML model development and internal validation (including 418 RCCs and 78 benign tumors), and patients from two independent institutions and a public database (TCIA) were included as the external dataset for individual testing (including 262 RCCs and 47 benign tumors). Features were extracted from three-phase CT images. CatBoost was used for feature selection and ML model establishment. The area under the receiver operating characteristic curve (AUC) was used to assess the performance of the ML model. The ML model based on 3D images performed better than that based on 2D images, with the highest AUC of 0.81 and accuracy (ACC) of 0.86. All three radiologists achieved better performance by referring to the classifier's decision, with accuracies increasing from 0.82 to 0.87, 0.82 to 0.88, and 0.76 to 0.87. The ML model achieved higher negative predictive values (NPV, 0.82-0.99), and the radiologists achieved higher positive predictive values (PPV, 0.91-0.95). A ML classifier based on whole-tumor three-phase CT images can be a useful and promising tool for differentiating RCCs from benign renal tumors. The ML model also perfectly complements radiologist interpretations. • A machine learning classifier based on CT images could be a reliable way to differentiate RCCs from benign renal tumors. • The machine learning model perfectly complemented the radiologists' interpretations. • Subtle variances in ROI delineation had little effect on the performance of the ML classifier.

Predicting future well performance for environmental remediation design using deep learning

In this study, we developed a deep learning (DL) framework with a multi-channel three-dimensional convolutional neural network (MC3D-CNN) to predict well performance and thereby assist future environmental remediation design. Such prediction of extraction well performance at designated locations is critical for configuring pump-and-treat (P&T) well network design and operation, setting reasonable target closure dates for overall remedying, and estimating remedy costs. The framework is developed with operational and monitoring data routinely collected during P&T remedy operations, including well extraction and injection rates as well as in situ contaminant concentrations. Traditionally, the collected data were rarely used for purposes other than assessing past well performance and the accuracy of the conceptual site model. However, recent advances in data-driven computational approaches enable better use of the large datasets to inform future well performance, enhance site characterization, and improve remediation planning. In this study, we established a DL framework to integrate transient three-dimensional contaminant plumes and multiple aquifer properties (e.g., hydraulic conductivity and hydrostratigraphic maps) to identify characteristic patterns controlling and representing extraction well mass recovery, aiming at providing future mass recovery estimates for existing wells and candidate wells at any proposed locations. We evaluated our framework by using a realistic synthetic dataset generated from a well-calibrated flow and transport model used in the 200 West Area of the U.S. Department of Energy’s Hanford Site in southeastern Washington state. The multi-channel feature in our framework allows integration of various types and temporal densities of training datasets for DL model development. Overall, we found that the trained DL model achieved an accuracy of over 90% in ranking extraction well performance in validation datasets, and over 80% in predicting high-performance-ranking well locations. This data-informed approach provides a flexible tool to support adaptive site management, streamline decision-making, and potentially reduce remediation time and costs. Our DL framework can be used as a filtering tool to improve the current P&T network optimization design by reducing the number of candidate well locations.

Early Pregnancy Blood Pressure Trajectories Predict Preeclampsia and Gestational Hypertension Among Low Risk Pregnancies

Develop and validate a prediction model for risk of preeclampsia (PE) and gestational hypertension (GH) using early pregnancy blood pressure (BP) patterns plus risk factors (RF) routinely available in clinical settings among low and moderate risk pregnancies. Retrospective cohort of 249,892 low or moderate risk singleton pregnancies delivered at Kaiser Permanente Northern California hospitals in 2009-2019. Electronic health records excluded prior hypertension, heart, kidney or liver disease and history of PE. All BPs were < 140/90 at < 20 weeks gestation. Data were randomly divided into a training dataset (N=174,925;70%) for model development and a validation dataset (n=74,967;30%). We used group-based trajectory modeling of longitudinal clinical systolic BPs to identify early pregnancy BP trajectory (BPT) groups < 20 weeks. We fit multinomial logistic regression models to predict early PE (< 34 weeks), late PE (≥34 weeks) and GH including BPT alone or with multiple RF (BMI, parity, age, race/ethnicity, diabetes) vs. an initial BP alone or with other RF. C-statistics generated from the validation dataset evaluated model predictive performance. The sample of 249,892 low to moderate risk individuals had mean (SD) maternal age of 30.9 y (5.3), BMI of 26.0 kg/m2 (5.8), 57% were nulliparous, and 63% were from Black, Hispanic, or Asian groups, and 100% had all normal BPs < 20 weeks. Overall, 992 (0.4%) developed early PE, 10,612 (4.3%) late PE, and 10,854 (4.3%) GH. BPT group models performed much better (C-statistic:0.664-0.732) than models with initial BP only (0.630-0.700), multiple RF alone, or in combination (Table 1). The best model performance (0.729-0.768) to predict early PE, late PE and GH included BPT groups plus BMI, parity, age, race/ethnicity, and diabetes. Six BP patterns plus RF available in routine clinical care improved prediction of early and late PE, and GH in lower risk pregnancies. Our findings show promise to guide surveillance and strategies to distinguish risk among ostensibly low risk groups.

A Review of Machine Learning Models for Harmful Algal Bloom Monitoring in Freshwater Systems

Highlights Machine Learning (ML) models are identified, reviewed, and analyzed for HAB predictions. Data preprocessing is vital for efficient ML model development. ML models for toxin production and monitoring are limited. Abstract. Harmful algal blooms (HABs) are detrimental to livestock, humans, pets, the environment, and the global economy, which calls for a robust approach to their management. While process-based models can inform practitioners about HAB enabling conditions, they have inherent limitations in accurately predicting harmful algal blooms. To address these limitations, Machine Learning (ML) models can potentially leverage large volumes of IoT data to aid in near real-time predictions. ML models have evolved as efficient tools for understanding patterns and relationships between water quality parameters and HAB expansion. This review describes ML models currently used for predicting and forecasting HABs in freshwater ecosystems and presents model structures and their application for predicting algal parameters and related toxins. The review revealed that regression trees, random forest, Artificial Neural Network (ANN), Support Vector Regression (SVR), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) are the most frequently used models for HABs monitoring. This review shows ML models' prowess in identifying significant variables influencing algal growth, HAB drivers, and multistep HAB prediction. Hybrid models also improve the prediction of algal-related parameters through improved optimization techniques and variable selection algorithms. While ML models often focus on algal biomass prediction, few studies apply ML models for toxin monitoring and prediction. This limitation can be associated with a lack of high-frequency toxin datasets for model development, and exploring this domain is encouraged. This review serves as a guide for policymakers and researchers to implement ML models for HAB prediction and reveals the potential of ML models for decision support and early prediction for HAB management. Keywords: Cyanobacteria, Freshwater, Harmful algal blooms, Machine learning, Water quality.