Confidence In Predictions Research Articles

Abstract 3132: CLUMPS-PTM: Spatial clustering of post-translational modifications across cancer types

Abstract Post-translational modifications (PTMs) play key roles in regulating cell signaling in health and disease processes. Recent advances in mass spectrometry enable comprehensive investigation of PTMs. However, identifying functionally relevant PTMs and protein domains implicated in disease remains a key challenge. We present CLUMPS-PTM, an algorithm for spatial clustering of PTMs to propose functional domains. We apply CLUMPS-PTM to the Clinical Proteomic Tumor Analysis Consortium (CPTAC) dataset (one of the largest proteogenomic pan-cancer PTM datasets) comprising 1110 samples spanning 11 tumor types. First, we map over 14K phospho-sites and 13K acetyl-sites detected in the CPTAC dataset to protein data bank (PDB) files. We expand this to the predicted AlphaFold proteome and recover additional 12K phospho-sites and 10K acetyl-sites exceeding model confidence (pLDDT) of 70%. Consistent with previous research, phospho-sites are found on residues with lower AlphaFold prediction confidence than acetyl-sites due to their abundance in unstructured domains (p-value < 1e-4). To identify functionally relevant protein domains in the pan-cancer context, we characterize the 1110 samples by applying SignatureAnalyzer, a Bayesian non-negative matrix factorization method, to discover multi-omic patterns. This analysis identifies strong signatures associated with DNA damage response (DDR) and stratifies the CPTAC cohort into DDR-high (n=442) and DDR-low (n=668) groups. Next, we applied CLUMPS-PTM to differentially modified phospho/acetyl-sites between these groups. We identified 22 proteins (including GMPS, NCBP1, and ANPEP (CD13)) with significant clusters of differentially phosphorylated sites (q-value < 0.1) in DDR-high samples. ANPEP is a zinc-dependent aminopeptidase with a significant cluster (q-value = 0.0129) of differentially phosphorylated sites in the ERAP1-C domain that faces peptidase active sites known to have proliferative activity via the MAPK pathway. CLUMPS-PTM also identified a significant cluster of acetyl-sites on ARID1A (a SWI/SNF chromatin remodeler known to be mutated in cancer) in DDR-high samples (q-value = 0.085). The ARID1A cluster is on the C-terminal tail of the protein within the glucocorticoid receptor (GR) domain, an upstream regulator of proteins that increase catabolism, reduce inflammation, and increase cell survival. Increased acetylation may block a ubiquitination site at the C-terminus of the protein, resulting in increased protein levels, likely due to reduced ARID1A degradation, and increased GR signaling. Overall, CLUMPS-PTM is an open-source tool that allows near proteome-wide spatial analysis with the growing availability of PTM data. We anticipate that it will be useful to the broader proteomic community for the discovery of novel targets and generation of insights into functional mechanisms. Citation Format: Shankara K. Anand, Yifat Geffen, Yo Akiyama, Francois Aguet, Gad Getz. CLUMPS-PTM: Spatial clustering of post-translational modifications across cancer types [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3132.

Abstract 1121: Patient-centric Discovery by Active Learning (PeDAL) platform to predict drug response in a heterogeneous patient population

Abstract Even after thorough preclinical testing, only 5% of oncology drug candidates entering clinical trials ultimately get approved for use in patients. To find one drug that will be effective in the clinic, drug developers have to test 20 agents, 19 of which will ultimately prove ineffective. Such high attrition rates drive up the cost of treatment and delay getting effective drugs to the market. Here we describe the Patient-centric Discovery by Active Learning (PeDAL) platform and present results of a study that shows it can predict the effectiveness of a drug or a set of drugs in a patient population with 92% accuracy in only 11 weeks while only performing in vitro assays on a fraction of the drug-patient combinations. The power of PeDAL is derived from a combination of AI technology, an extensive drug response and biomarker database, and a biobank of 150,000 tumor samples from 137 tumor types. The study consists of two phases, the first demonstrates that PeDAL’s predictive model efficiently reveals drug-response patterns that provide insight into the treatment of ovarian cancers. The second phase focuses on the in vitro validation of the accuracy of the predictions. The experimental space was 175 FDA-approved drugs and 130 primary ovarian tumor samples. PeDAL’s active machine learning algorithm iteratively optimizes predictive models by learning from in vitro experiments which take into account tumor-stromal interactions and cellular heterogeneity. Importantly, the ability to model patient-drug combinations across a diverse population of primary samples adds a patient heterogeneity dimension to the generated drug response predictions. After performing 720 high throughput experiments, PeDAL was able to make drug-response predictions for an additional 4,600 drug/tumor predictions. Additional in vitro validation experiments confirm the reproducibility and accuracy of the high-confidence predictions generated by PeDAL with area under the curve (AUC) values >0.95. These results have strong implications not only for the new drug discovery process, but also for identifying opportunities for drug repurposing. We demonstrate that the PeDAL platform is a highly efficient tool to make confident predictions about drug-tumor pairings allowing the discovery of many more drug-tumor responses than are practical to test in vitro. Citation Format: Robert J. Montgomery, Aaron W. Gilkey, Joshua D. Kangas, Robert F. Murphy, Arlette H. Uihlein. Patient-centric Discovery by Active Learning (PeDAL) platform to predict drug response in a heterogeneous patient population [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 1121.

Sustaining the High Performance of AI-Based Network Traffic Classification Models

Network traffic classification plays an essential role in network measurement and management. Emerging Artificial Intelligence (AI) algorithms have become a viable solution to encrypted network traffic classification. Nonetheless, the classification performance of existing AI-based traffic classifiers is restricted to a limited number of network applications depending on the coverage of the knowledge database. Such AI-based traffic classifiers cannot maintain high performance to provide accurate traffic classification when dealing with updated or new network applications. To tackle the issues, we present an autonomous model update mechanism to sustain the high performance of AI-based traffic classifiers. Specifically, an instability check algorithm is derived to evaluate if the current classifier requires an update. A filtering algorithm is proposed to extract unknown traffic and build a new knowledge database based on a new metric, i.e., familiarity, defined based on the prediction confidence and instability. Extensive experiment results demonstrate that our proposed updating mechanism can provide prompt model updates and establish a proper new knowledge base to maintain high accuracy in various experimental scenarios. Moreover, the comparison is conducted and the results show the proposed familiarity-based filtering algorithm can filter about 7 and 3 times more true positive packets in the two considered scenarios, respectively.

Effects of resistance training on body weight and body composition in older adults: An inter-individual response difference meta-analysis of randomized controlled trials.

Whether true inter-individual response differences (IIRD) occur as a result of resistance training on body weight and body composition in older adults with overweight and obesity is not known. To address this gap, data from a previous meta-analysis representing 587 men and women (333 resistance training, 254 control) ≥ 60 years of age nested in 15 randomized controlled trials of resistance training ≥ 8 weeks were included. Resistance training and control group change outcome standard deviations treated as point estimates for body weight and body composition (percent body fat, fat mass, body mass index in kg.m2, and lean body mass) were used to calculate true IIRD from each study. True IIRD as well as traditional pairwise comparisons were pooled using the inverse-variance (IVhet) model. Both 95% confidence intervals (CI) and prediction intervals (PI) were calculated. While statistically significant improvements were found for body weight and all body composition outcomes (p < 0.05 for all), no statistically significant IIRD was observed for any of the outcomes (p > 0.05 for all) and all 95% PIs overlapped. Conclusions: While resistance training is associated with improvements in body weight and body composition in older adults, the lack of true IIRD suggests that factors other than training response variation (random variation, physiological responses associated with behavioral changes that are not the result of resistance training) are responsible for the observed variation in body weight and body composition.

PD22-07 A MULTICENTER STUDY PRESENTING AN EXPLAINABLE AI APPROACH TO AUTOMATIC SEGMENTATION OF RENAL TUMORS ON COMPUTED TOMOGRAPHY

You have accessJournal of UrologyCME1 Apr 2023PD22-07 A MULTICENTER STUDY PRESENTING AN EXPLAINABLE AI APPROACH TO AUTOMATIC SEGMENTATION OF RENAL TUMORS ON COMPUTED TOMOGRAPHY Annemarie Uhlig, Sophie Bachanek, Paul Romario Würzberg, Manuel Nietert, Lutz Trojan, Joachim Lotz, and Johannes Uhlig Annemarie UhligAnnemarie Uhlig More articles by this author , Sophie BachanekSophie Bachanek More articles by this author , Paul Romario WürzbergPaul Romario Würzberg More articles by this author , Manuel NietertManuel Nietert More articles by this author , Lutz TrojanLutz Trojan More articles by this author , Joachim LotzJoachim Lotz More articles by this author , and Johannes UhligJohannes Uhlig More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003295.07AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Automatic segmentation of renal tumors is a cornerstone for Artificial Intelligence (AI) based tumor diagnostics. This study presents an automatic segmentation algorithm and visualization method using CT-studies from renal tumor patients recruited in a multicenter setting. METHODS: Renal tumor patients diagnosed between 2018 and 2021 were retrospectively assessed for this study. Inclusion criteria were CT-imaging of renal tumors in corticomedullary (CM) or nephrogenic (NG) contrast media phase. Patients aged <18years, and those with cystic or infiltrative renal tumors (i.e. lymphomas) were excluded. Manual segmentation of the renal tumors was performed by a GU-radiologist on all axial CT slices. A convolutional neural network (UNET) was trained based on the radiologists´ manual segmentations. In an independent validation dataset, the accuracy of the UNETs predictions of renal tumor contours was compared to the references standard of manual segmentations and quantified using the DICE score. RESULTS: A total of n=394/ n=350 patients with renal tumors imaged in CM /NG phase, respectively, were included (median age 66y; 35% female; median tumor diameter 5.4cm). CT-studies from >20 radiological imaging centers were included with different imaging protocols and slice thickness. The UNET was trained on n=316 CM and n=294 NG contrast phase patients (n=7019 / n=6859 separate CT images). In the independent validation dataset (n=78 / n=56 patients with 1713 / 1298 CT images), the UNET achieved a DICE score of 0.88 and 0.90 for the corticomedullary and nephrogenic CM phase, respectively. The UNET predictions were visualized using a tile-based approach with color-coding and contour-lines that could be overlaid on CT-images to depict varying levels of prediction confidence. CONCLUSIONS: A UNET AI-approach yields a robust automatic delineation of renal tumors on CT-images acquired in clinical routine, irrespective of the contrast media phase. Using color-coding and contour-lines that could be overlaid on original CT-images provides an explainable approach to the UNETs predictions and might improve AI-acceptance in clinical practice. Source of Funding: This study received funding by the 2022 ESR Research Seed Grant. The 2022 ESR Research Seed Grants were kindly supported by an unrestricted, non-exclusive grant from GE Healthcare © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e667 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Annemarie Uhlig More articles by this author Sophie Bachanek More articles by this author Paul Romario Würzberg More articles by this author Manuel Nietert More articles by this author Lutz Trojan More articles by this author Joachim Lotz More articles by this author Johannes Uhlig More articles by this author Expand All Advertisement PDF downloadLoading ...

PD34-05 A NOVEL MACHINE-LEARNING ALGORITHM TO PREDICT STONE RECURRENCE WITH 24-HOUR URINE DATA

You have accessJournal of UrologyCME1 Apr 2023PD34-05 A NOVEL MACHINE-LEARNING ALGORITHM TO PREDICT STONE RECURRENCE WITH 24-HOUR URINE DATA Kevin Shee, Carter Chan, Andrew Liu, Heiko Yang, Wilson Sui, Sunita Ho, Thomas Chi, and Marshall Stoller Kevin SheeKevin Shee More articles by this author , Carter ChanCarter Chan More articles by this author , Andrew LiuAndrew Liu More articles by this author , Heiko YangHeiko Yang More articles by this author , Wilson SuiWilson Sui More articles by this author , Sunita HoSunita Ho More articles by this author , Thomas ChiThomas Chi More articles by this author , and Marshall StollerMarshall Stoller More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003327.05AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: There are limited predictive markers of stone recurrence, which poses a challenge for the clinical management of stone disease. Furthermore, inability to predict stone events is a significant limitation for clinical trials, where many patients must be enrolled to obtain sufficient stone events for analysis. In this study, we sought to use machine learning methods to identify a novel algorithm to predict stone recurrence. METHODS: Patients enrolled in the Registry for Stones of the Kidney and Ureter (ReSKU), a multi-institutional registry of nephrolithiasis patients collected between 2015-2020, with at least one prospectively collected 24-hour urine test (Litholink) were included in the training set. A test set of stone patients not enrolled in ReSKU with 24-hour urine test results were collected. Stone events were defined as either an office visit where a patient reported symptomatic passage of stone(s) or surgery for stone removal. Seven prediction classification methods were evaluated: Decision Tree, Logistic regression with ElasticNet, Extra Trees, kNeighbors, LightGBM, Logistic Regression, and Random Forest. Probability of Recurrent Nephrolithiasis (PRN) score ranging from 0-1 was generated from the models. Predictive analyses and ROC curve generation was performed in R. RESULTS: A training set of 423 kidney stone patients with stone event data and 24-hour urine samples were trained using the prediction classification methods. The highest performing prediction model was a Logistic Regression with ElasticNet machine learning model (AUC=0.65; Figure 1A). Selecting for high confidence predictions (PRN 0-0.15, 0.8-1.0) improved model accuracy (AUC=0.82; Figure 1B). The model was validated on a test set of 172 stone patients. Patients predicted to have high likelihood of stone recurrence had an increase in the average number of stone events compared to those with low likelihood (3.6 vs 1.7 events; p=2e-7). Prediction accuracy in the test set demonstrated moderate discriminative ability (AUC=0.64; Figure 1C). CONCLUSIONS: Using machine-learning trained on 24-hour urine and stone outcome data, a novel algorithm was developed to predict stone recurrence. Initial validation studies suggest ability of algorithm to identify patients with higher likelihood of stone events and warrants further study. Source of Funding: n/a © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e921 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Kevin Shee More articles by this author Carter Chan More articles by this author Andrew Liu More articles by this author Heiko Yang More articles by this author Wilson Sui More articles by this author Sunita Ho More articles by this author Thomas Chi More articles by this author Marshall Stoller More articles by this author Expand All Advertisement PDF downloadLoading ...

An Efficient and Robust Cloud-Based Deep Learning With Knowledge Distillation

In recent years, deep neural networks have shown extraordinary power in various practical learning tasks, especially in object detection, classification, natural language processing. However, deploying such large models on resource-constrained devices or embedded systems is challenging due to their high computational cost. Efforts such as model partition, pruning, or quantization have been used at the expense of accuracy loss. Knowledge distillation is a technique that transfers model knowledge from a well-trained model (teacher) to a smaller and shallow model (student). Instead of using a learning model on the cloud, we can deploy distilled models on various edge devices, significantly reducing the computational cost, memory usage and prolonging the battery lifetime. In this work, we propose a novel neuron manifold distillation (NMD) method, where the student models imitate the teacher's output distribution and learn the feature geometry of the teacher model. In addition, to further improve the cloud-based learning system reliability, we propose a confident prediction mechanism to calibrate the model predictions. We conduct experiments with different distillation configurations over multiple datasets. Our proposed method demonstrates a consistent improvement in accuracy-speed trade-offs for the distilled model.

Information Entropy Measures for Evaluation of Reliability of Deep Neural Network Results.

Deep neural networks (DNN) try to analyze given data, to come up with decisions regarding the inputs. The decision-making process of the DNN model is not entirely transparent. The confidence of the model predictions on new data fed into the network can vary. We address the question of certainty of decision making and adequacy of information capturing by DNN models during this process of decision-making. We introduce a measure called certainty index, which is based on the outputs in the most penultimate layer of DNN. In this approach, we employed iEEG (intracranial electroencephalogram) data to train and test DNN. When arriving at model predictions, the contribution of the entire information content of the input may be important. We explored the relationship between the certainty of DNN predictions and information content of the signal by estimating the sample entropy and using a heatmap of the signal. While it can be assumed that the entire sample must be utilized for arriving at the most appropriate decisions, an evaluation of DNNs from this standpoint has not been reported. We demonstrate that the robustness of the relationship between certainty index with the sample entropy, demonstrated through sample entropy-heatmap correlation, is higher than that with the original signal, indicating that the DNN focuses on information rich regions of the signal to arrive at decisions. Therefore, it can be concluded that the certainty of a decision is related to the DNN's ability to capture the information in the original signal. Our results indicate that, within its limitations, the certainty index can be used as useful tool in estimating the confidence of predictions. The certainty index appears to be related to how effectively DNN heatmaps captured the information content in the signal.

Fluid-structure interaction simulation for performance prediction and design optimization of parafoils

Parachute design is challenging to achieve innovative progress if the dominant role of testing continues, as it will be an increasingly expensive and time-consuming work. The aim of this study is to establish a reliable and efficient design tool using existing advanced numerical modeling methods. This paper presents a numerical method based on two-way coupled fluid-structure interaction (FSI) strategies for predicting aerodynamic and flight performance for parafoil design optimization. The nonlinear finite element method was used for the canopy fabric model and flow field, and the fluid dynamics were solved by Reynolds-averaged Navier-Stokes with the Spalart-Allmaras turbulence model. The FSI simulations are performed to assess the aerodynamic performance and structural deformations of full-scale parafoil canopies. The equilibrium shape of the parafoil canopy under steady gliding states and the relevant flow field were analyzed to enhance confidence and understanding in the performance prediction of new parachutes. Three-dimensional FSI simulation results of parafoils show that the inflation caused flexible bulges of canopy cells, and the maximum lift coefficient increased more than 16% with a higher stall angle of attack than that of the rigid body model. A parafoil with a smaller leading edge inlet or a scaling down area can improve the aerodynamic performance, mainly manifested in a higher lift-to-drag ratio and better anti-stall performance. Finally, the prediction results of parafoil glide performance were verified by flight test data, and the prediction accuracy of the flexible model is more than 10% higher than that of the rigid model. This work makes the simulation tools a step closer to practical application.

Toward reliable calcification detection: calibration of uncertainty in object detection from coronary optical coherence tomography images.

Optical coherence tomography (OCT) has become increasingly essential in assisting the treatment of coronary artery disease (CAD). However, unidentified calcified regions within a narrowed artery could impair the outcome of the treatment. Fast and objective identification is paramount to automatically procuring accurate readings on calcifications within the artery. We aim to rapidly identify calcification in coronary OCT images using a bounding box and reduce the prediction bias in automated prediction models. We first adopt a deep learning-based object detection model to rapidly draw the calcified region from coronary OCT images using a bounding box. We measure the uncertainty of predictions based on the expected calibration errors, thus assessing the certainty level of detection results. To calibrate confidence scores of predictions, we implement dependent logistic calibration using each detection result's confidence and center coordinates. We implemented an object detection module to draw the boundary of the calcified region at a rate of 140 frames per second. With the calibrated confidence score of each prediction, we lower the uncertainty of predictions in calcification detection and eliminate the estimation bias from various object detection methods. The calibrated confidence of prediction results in a confidence error of , suggesting that the confidence calibration on calcification detection could provide a more trustworthy result. Given the rapid detection and effective calibration of the proposed work, we expect that it can assist in clinical evaluation of treating the CAD during the imaging-guided procedure.

Finite element simulation of growth modulation during brace treatment of adolescent idiopathic scoliosis.

Adolescent idiopathic scoliosis (AIS) is a spine deformity whose progression during growth is affected by asymmetrical loads acting on the spine. The conservative brace treatment aims to limit the deformity's progression until the end of skeletal growth. This study's objective was to develop a patient-specific finite element model (FEM) simulating immediate in-brace correction and subsequent growth modulation over 2 years of treatment. Thirty-five retrospective AIS cases with documented correction over 2 years were analyzed. For each case, a patient-specific FEM was built, and in-brace correction was simulated. Vertebral growth and its modulation were modeled using simulated pressures on epiphyseal vertebral growth plates, including a compliance factor representing the recorded brace wear. The simulated Cobb angles, thoracic kyphosis, lumbar lordosis and apical vertebral rotation were compared with the actual measurements immediately in-brace and out-of-brace at the 2-year follow-up. Treatment outcomes according to simulated compliance scenarios of no brace-wear vs. full brace-wear were also computed. The average immediate in-brace difference between the simulated and actual Cobb angle was 4.9° (MT) and 3.7° (TL/L). 2-year out-of-brace, it was 5.6° (MT) and 5.4° (TL/L). The no brace-wear and full brace-wear compliance scenarios resulted respectively in 15/35 (43%) and 31/35 (89%) simulated spine deformities progressing by less than 5° over 2 years of treatment. Clinical significance: the FEM's ability to simulate the final correction with an accuracy on the order of the radiological measurements' interoperator reproducibility, combined with its sensitivity to brace-wear compliance, provides confidence in the model's predictions for a comparative context of use like improving a brace's design before its application. This article is protected by copyright. All rights reserved.

Vulvovaginal yeast infections during pregnancy and perinatal outcomes: systematic review and meta-analysis

BackgroundVulvovaginal yeast infections in pregnancy are common and can cause extensive inflammation, which could contribute to adverse pregnancy outcomes. Symptomatic yeast infections are likely to cause more inflammation than asymptomatic. The objective of this study was to investigate associations between symptomatic and asymptomatic vulvovaginal yeast infections in pregnancy and perinatal outcomes.MethodsWe did a systematic review and searched eight databases until 01 July 2022. We included studies reporting on pregnant women with and without laboratory confirmed vulvovaginal yeast infection and preterm birth or eight other perinatal outcomes. We used random effects meta-analysis to calculate summary odds ratios (OR), 95% confidence intervals (CI) and prediction intervals for the association between yeast infection and outcomes. We described findings from studies with multivariable analyses. We assessed the risk of bias using published tools.ResultsWe screened 3909 references and included 57 studies. Only 22/57 studies reported information about participant vulvovaginal symptoms. Preterm birth was an outcome in 35/57 studies (49,161 women). In 32/35 studies with available data, the summary OR from univariable analyses was 1.01 (95% CI 0.84–1.21, I2 60%, prediction interval 0.45–2.23). In analyses stratified by symptom status, we found ORs of 1.44 (95% CI 0.92–2.26) in two studies with ≥ 50% symptomatic participants, 0.84 (95% CI 0.45–1.58) in seven studies with < 50% symptomatic participants, and 1.12 (95% CI 0.94–1.35) in four studies with asymptomatic participants. In three studies with multivariable analysis, adjusted ORs were greater than one but CIs were compatible with there being no association. We did not find associations between vulvovaginal yeast infection and any secondary outcome. Most studies were at high risk of bias in at least one domain and only three studies controlled for confounding.ConclusionsWe did not find strong statistical evidence of an increased risk for preterm birth or eight other adverse perinatal outcomes, in pregnant women with either symptomatic or asymptomatic vulvovaginal yeast infection. The available evidence is insufficient to make recommendations about testing and treatment of vulvovaginal yeast infection in pregnancy. Future studies should assess vulvovaginal symptoms, yeast organism loads, concomitant vaginal or cervical infections, and microbiota using state-of-the-art diagnostics.Systematic review registrationPROSPERO CRD42020197564

Quantifying uncertainty in the temporal disposition of groundwater inundation under sea level rise projections

Over the next century, coastal regions are under threat from projected rising sea levels and the potential emergence of groundwater at the land surface (groundwater inundation). The potential economic and social damages of this largely unseen, and often poorly characterised natural hazard are substantial. To support risk-based decision making in response to this emerging hazard, we present a Bayesian modelling framework (or workflow), which maps the spatial distribution of groundwater level uncertainty and inundation under Intergovernmental Panel on Climate Change (IPCC) projections of Sea Level Rise (SLR). Such probabilistic mapping assessments, which explicitly acknowledge the spatial uncertainty of groundwater flow model predictions, and the deep uncertainty of the IPCC-SLR projections themselves, remains challenging for coastal groundwater systems. Our study, therefore, presents a generalisable workflow to support decision makers, that we demonstrate for a case study of a low-lying coastal region in Aotearoa New Zealand. Our results provide posterior predictive distributions of groundwater levels to map susceptibility to the groundwater inundation hazard, according to exceedance of specified model top elevations. We also explore the value of history matching (model calibration) in the context of reducing predictive uncertainty, and the benefits of predicting changes (rather than absolute values) in relation to a decision threshold. The latter may have profound implications for the many at-risk coastal communities and ecosystems, which are typically data poor. We conclude that history matching can indeed increase the spatial confidence of posterior groundwater inundation predictions for the 2030-2050 timeframe.

SemiRoadExNet: A semi-supervised network for road extraction from remote sensing imagery via adversarial learning

Road extraction from remote sensing imagery is a popular and frontier research focus, since road information plays an essential role in application fields, such as urban management, map updating and traffic planning. Deep learning based methods have shown their dominance on road extraction from remote sensing imagery. However, the performance of the existing road extraction methods relies heavily on a large amount of high-quality annotated training data, which is usually hard to obtain in practice. Current semi-supervised road extraction models can effectively reduce the dependency on the labeled data, nevertheless they cannot fully utilize the latent information of low-confidence pixels in pseudo-labels effectively. These pixels are usually a border between road and non-road area and of importance for the prediction accuracy of road extraction models. In order to address these issues, we proposed a novel semi-supervised road extraction network, named SemiRoadExNet. SemiRoadExNet is based on a Generative Adversarial Network (GAN), containing one generator with two discriminators. Firstly, both labeled and unlabeled images are put into the generator network for road extraction, and the outputs of the generator not only include road segmentation results but also the corresponding entropy maps. The entropy maps represent the confidence of prediction (road or non-road) for each pixel. Then, the two discriminators enforce the feature distributions keeping the consistency of road prediction maps and entropy maps between the labeled and unlabeled data. During the adversarial training, the generator is continuously regularized by exploiting the potential information from unlabeled data, thus the generalization capacity of the proposed model can be improved effectively. Compared to several state-of-the-art semi-supervised semantic segmentation methods, the proposed SemiRoadExNet achieves 0.96–5.38% IoU improvements on DeepGlobe Road Extraction, Massachusetts Roads and CHN6-CUG datasets respectively. The source code of d SemiRoadExNet is freely available at https://github.com/hchen118/SemiRoadExNet.

TCNN: A Transformer Convolutional Neural Network for artifact classification in whole slide images

The production of pathological slides is a complex task requiring several physical and chemical procedures that are often done manually. Occasionally, such procedures may end up in patterns (called artifacts) appearing in the images that were not present in the biological tissue but are produced during the slide processing. This paper presents a new method for the automatic detection of such artifacts in pathological images. This detection is seen as a binary classification task that is solved by means of a new Transformer Convolutional Neural Network (TCNN). The proposed method can assist laboratory technicians in avoiding the time-consuming manual labeling of images, preventing the risk of sending artifact patches to pathologists and physicians. The artifact patches include misleading data and should not be considered. In addition, artifact patches jeopardize the accuracy of Computer-Aided Diagnosis (CAD) systems. The proposed method works on the Hue-Saturation-Value colormap layer, which is related to color representation, to improve model robustness against external lighting changes. Spatial Transformer layers are used to make the model spatially invariant on input data, and a set of convolutional layers of EfficientNet Convolutional Neural Network improve pattern recognition. Images used in this research are more than 13,000 and represent 22 types of tissue affected by 7 diverse types of artifacts. Through performance measurement, we evaluate the model using several metrics: Probability Calibration graph, Mean Prediction Probability, Brier Loss, ROC-AUC curve, Confusion Matrix, Accuracy, Sensitivity, Specificity, F1-Score, and Confidence Interval.

Clinicians' Prediction of Survival and Prognostic Confidence in Patients with Advanced Cancer in Three East Asian Countries.

Background: Little is known about accuracy and confidence of clinicians' prediction of survival (CPS) in East-Asian countries. Objective: We aimed to examine accuracy of CPS for 7-, 21-, and 42-day survival in palliative inpatients and its association with prognostic confidence. Design: An international prospective cohort study in Japan (JP), Korea (KR), and Taiwan (TW). Setting/Subjects: Subjects were inpatients with advanced cancer admitted to 37 palliative care units in three countries. Measurements: Discrimination of CPS was investigated through sensitivity, specificity, overall accuracy, and area under the receiver operating characteristics curves (AUROCs) according to 7-, 21-, and 42-day survival. The accuracies of CPS were compared with those of Performance Status-based Palliative Prognostic Index (PS-PPI). Clinicians were instructed to rate confidence level on a 0-10-point scale. Results: A total of 2571 patients were analyzed. The specificity was highest at 93.2-100.0% for the 7-day CPS, and sensitivity was highest at 71.5-86.8% for the 42-day CPS. The AUROCs of the seven-day CPS were 0.88, 0.94, and 0.89, while those of PS-PPI were 0.77, 0.69, and 0.69 for JP, KR, and TW, respectively. As for 42-day prediction, sensitivities of PS-PPI were higher than those of CPS. Clinicians' confidence was strongly associated with the accuracy of prediction in all three countries (all p-values <0.01). Conclusions: CPS accuracies were highest (0.88-0.94) for the seven-day survival prediction. CPS was more accurate than PS-PPI in all timeframe prediction except 42-day prediction in KR. Prognostic confidence was significantly associated with the accuracy of CPS.

Using serial dependence to predict confidence across observers and cognitive domains.

Our perceptual system appears hardwired to exploit regularities of input features across space and time in seemingly stable environments. This can lead to serial dependence effects whereby recent perceptual representations bias current perception. Serial dependence has also been demonstrated for more abstract representations, such as perceptual confidence. Here, we ask whether temporal patterns in the generation of confidence judgments across trials generalize across observers and different cognitive domains. Data from the Confidence Database across perceptual, memory, and cognitive paradigms was reanalyzed. Machine learning classifiers were used to predict the confidence on the current trial based on the history of confidence judgments on the previous trials. Cross-observer and cross-domain decoding results showed that a model trained to predict confidence in the perceptual domain generalized across observers to predict confidence across the different cognitive domains. The recent history of confidence was the most critical factor. The history of accuracy or Type 1 reaction time alone, or in combination with confidence, did not improve the prediction of the current confidence. We also observed that confidence predictions generalized across correct and incorrect trials, indicating that serial dependence effects in confidence generation are uncoupled to metacognition (i.e., how we evaluate the precision of our own behavior). We discuss the ramifications of these findings for the ongoing debate on domain-generality versus domain-specificity of metacognition.

Prevalence of grief symptoms and disorders in the time of COVID-19 pandemic: A meta-analysis.

Millions of people worldwide are mourning the loss of loved ones due to the COVID-19 epidemic, which may adversely impact their mental health. This meta-analysis aimed to investigate pandemic grief symptoms and disorders for developing policy, practice, and research priorities. The databases of Cochrane, Embase, Ovid-MEDLINE, WHO COVID-19, NCBI SARS-CoV-2, Scopus, Web of Science, CINAHL, and Science Direct were comprehensively searched until July 31, 2022. The Joanna Briggs Institute's and Hoy's criteria were used to evaluate the studies. A pooled prevalence was presented in a forest plot figure with a corresponding 95% confidence interval (CI) and prediction interval. Between-study heterogeneity was measured using the I2 and Q statistics. Variations in the prevalence estimates in different subgroups were examined by moderator meta-analysis. The search identified 3677 citations, of which 15 studies involving 9289 participants were included in the meta-analysis. The pooled prevalence rate of grief symptoms was 45.1% (95%CI: 32.6%-57.5%), and grief disorder was 46.4% (95%CI: 37.4%-55.5%). Grief symptoms were significantly higher in <6 months (45.8%; 95%CI: 26.3%-65.3%) compared to the period of more than 6 months. Unfortunately, moderator analyses for grief disorders could not be performed due to limited studies. The prevalence of grief problems during the pandemic was substantially higher than in non-pandemic conditions; therefore, it is crucial to strengthen bereavement support to reduce psychological distress. The results provide a foundation for nurses and healthcare workers to anticipate a heightened need for support and provision of bereavement care in the post-pandemic era.

Urban Flood Modeling: Uncertainty Quantification and Physics‐Informed Gaussian Processes Regression Forecasting

AbstractEstimating uncertainty in flood model predictions is important for many applications, including risk assessment and flood forecasting. We focus on uncertainty in physics‐based urban flooding models. We consider the effects of the model's complexity and uncertainty in key input parameters. The effect of rainfall intensity on the uncertainty in water depth predictions is also studied. As a test study, we choose the Interconnected Channel and Pond Routing (ICPR) model of a part of the city of Minneapolis. The uncertainty in the ICPR model's predictions of the floodwater depth is quantified in terms of the ensemble variance using the multilevel Monte Carlo (MC) simulation method. Our results show that uncertainties in the studied domain are highly localized. Model simplifications, such as disregarding the groundwater flow, lead to overly confident predictions, that is, predictions that are both less accurate and uncertain than those of the more complex model. We find that for the same number of uncertain parameters, increasing the model resolution reduces uncertainty in the model predictions (and increases the MC method's computational cost). We employ the multilevel MC method to reduce the cost of estimating uncertainty in a high‐resolution ICPR model. Finally, we use the ensemble estimates of the mean and covariance of the flood depth for real‐time flood depth forecasting using the physics‐informed Gaussian process regression method. We show that even with few measurements, the proposed framework results in a more accurate forecast than that provided by the mean prediction of the ICPR model.

Preincubation Time-Dependent, Long-Lasting Inhibition of Drug Transporters and Impact on the Prediction of Drug-Drug Interactions.

Transporter-mediated drug-drug interaction (DDI) is of clinical concern, and the quantitative prediction of DDIs is an indispensable part of drug development. Cell-based inhibition assays, in which a representative probe substrate and a potential inhibitor are coincubated, are routinely performed to assess the inhibitory potential of new molecular entities on drug transporters. However, the inhibitory effect of cyclosporine A (CsA) on organic anion transporting polypeptide (OATP) 1B1 is substantially potentiated with CsA preincubation, and this effect is both long-lasting and dependent on the preincubation time. This phenomenon has also been reported with transporters other than OATP1Bs, but it is considered more prevalent among OATP1Bs and organic cation transporters. Regulatory agencies have also noted this preincubation effect and have recommended that pharmaceutical companies consider inhibitor preincubation when performing in vitro OATP1B1 and OATP1B3 inhibition studies. Although the underlying mechanisms responsible for the preincubation effect are not fully understood, a trans-inhibition mechanism was recently demonstrated for OATP1B1 inhibition by CsA, in which CsA inhibited OATP1B1 not only extracellularly (cis-inhibition) but also intracellularly (trans-inhibition). Furthermore, the trans-inhibition potency of CsA was much greater than that of cis-inhibition, suggesting that trans-inhibition might be a key driver of clinical DDIs of CsA with OATP1B substrate drugs. Although confidence in transporter-mediated DDI prediction is generally considered to be low, the predictability might be further improved by incorporating the trans-inhibition mechanism into static and dynamic models for preincubation-dependent inhibitors of OATP1Bs and perhaps other transporters. SIGNIFICANCE STATEMENT: Preincubation time-dependent, long-lasting inhibition has been observed for OATP1B1 and other solute carrier transporters in vitro. Recently, a trans-inhibition mechanism for the preincubation effect of CsA on OATP1B1 inhibition was identified, with the trans-inhibition potency being greater than that of cis-inhibition. The concept of trans-inhibition may allow us to further understand the mechanism of transporter-mediated DDIs not only for OATP1B1 but also for other transporters and to improve the accuracy and confidence of DDI predictions.