Flare Prediction Research Articles

The development of PIL extraction application in active region of SDO/HMI magnetogram

The Polarity Inversion Line (PIL) on the active region (AR) is a parameter that correlates with the occurrence of the solar flare. This research has developed application to extract the PIL length in the AR. Image processing techniques were used to detect and classify AR from the SDO/HMI Magnetogram image of 1024x1024 pixels. The detection of AR in magnetogram images has its challenges because the ARs are constructed by the white (black) regions represent the magnetic positive (negative) polarities. To improve detection accuracy, the application used The Solar Region Summary (SRS), compiled by SWPC, and the centroid position of sunspot on SDO/HMI Continuum as additional information. The results of AR detection had high accuracy because SRS stores the centroid position of the flaring active region at peak time with the Stonyhurst heliographic on the solar disk which is converted to Cartesian coordinates. The application test used 1450 images and obtained the PIL length on the peak time of the X-class and M-class flares from 2011-2017. The application detected the high distribution of PIL length in the range of 120 to 580 Mm for X class flare, and 120 to 390 Mm for M-class flare. This research provides new data that has the potential to improve the machine learning-based solar flare prediction models.

Risk factors of flare in patients with systemic lupus erythematosus after glucocorticoids withdrawal. A systematic review and meta-analysis.

ObjectiveGlucocorticoids (GC) withdrawal is part of the targets in current recommendations for SLE, but relapse is the most worrying issue. We aimed to investigate the predictors for flare in patients with SLE after GC withdrawal.MethodsWe systematically searched PubMed, EMBASE and Cochrane Library as well as Scopus databases up to 9 July 2021 for studies concerning predictive factors of relapses in patients with SLE after GC cessation. Pooled OR and 95% CI were combined using a random-effects or fixed-effects model.Results635 patients with SLE with GC discontinuation in 9 publications were eligible for the final analysis. Of them, 99.5% patients were in clinical remission before GC withdrawal. Serologically active yet clinically quiescent (SACQ) was associated with an increased risk of flare after GC withdrawal (OR 1.78, 95% CI (1.00 to 3.15)). Older age and concomitant use of hydroxychloroquine (HCQ) trended towards decreased risk of flare (weighted mean difference (WMD) −2.04, 95% CI (−4.15 to 0.06) for age and OR 0.50, 95% CI (0.23 to 1.07) for HCQ), yet not statistically significant. No significant association was observed regarding gender (pooled OR 1.75; 95% CI (0.59 to 5.20)), disease duration (WMD −11.91, 95% CI (−27.73 to 3.91)), remission duration (WMD −8.55, 95% CI (−33.33 to 16.23)), GC treatment duration (WMD −10.10, 95% CI (−64.09 to 43.88)), concomitant use of immunosuppressant (OR 0.86, 95% CI (0.48 to 1.53)).ConclusionYounger age and SACQ were potential risk factors of SLE flare among patients who discontinued GC. HCQ, but not immunosuppressant might prevent flare. GC withdrawal should be done with caution in this subgroup of patients.

Hydroxychloroquine blood levels predicts flare in childhood-onset lupus nephritis.

Low hydroxychloroquine (HCQ) blood levels are predictors of flare in adult lupus. Childhood-onset systemic lupus erythematosus (cSLE) has high morbidity with renal involvement in up to 80% of cases. The aim of this study is to determine the HCQ cut-off levels which predicts flare in childhood-onset lupus nephritis (LN). Sixty LN patients on HCQ use for at least 6-months were prospectively evaluated at baseline (BL) and about 6-months later for cSLE flare and HCQ blood levels (ng/mL) measured by liquid chromatography-tandem mass spectrometry. There were 19 patients (32%) with flare, during the study with median SLEDAI increase of 4 (0-8). Median (IQR) BL HCQ levels of the flare group were lower compared to stable patients [557.5 (68.6-980.3) vs. 1061.9 (534.8-1590.0ng/mL); p=0.012]. ROC curve analysis demonstrated that HCQ levels≤1075ng/mL were associated with a 5.08 (95%CI 1.28-20.13; p=0.021) times increased risk of flare. Six-month HCQ levels revealed that most patients 24/54 (44%) had persistently low levels (≤1075) during follow-up. Among those, 11/24 (46%) had flare. Multiple logistic regression analysis including prednisone use, baseline SLEDAI-2K, adherence based on pharmacy refill and BL HCQ blood levels as possible predictors of flare revealed that only HCQ blood level was independently associated with flare (OR 0.999, 95%CI 0.998-1.0, p=0.013). We demonstrated that HCQ blood cut-off level under 1075ng/mL predicts flare in childhood-onset LN patients under prescribed HCQ dose of 4.0-5.5mg/kg/day. We further observed that most of these patients have compliance issues reinforcing the need for a close surveillance particularly in those with levels below the defined cut-off.

Solar Flare Prediction Based on the Fusion of Multiple Deep-learning Models

Abstract Solar flare formation mechanisms and their corresponding predictions have commonly been difficult topics in solar physics for decades. The traditional forecasting method manually constructs a statistical relationship between the measured values of solar active regions and solar flares that cannot fully utilize the information related to solar flares contained in observational data. In this article, we first used neural-network methods driven by the measured magnetogram and magnetic characteristic parameters of the sunspot group to learn the prediction model and predict solar flares. The prediction fusion model is based on a deep neural network, convolutional neural network, and bidirectional long short-term memory neural network and can predict whether a sunspot group will have a flare event above class M or class C in the next 24 or 48 hr. The real skill statistics (TSS) and F1 scores were used to evaluate the performances of our fusion model. The test results clearly show that this fusion model can make full use of the information related to solar flares and combine the advantages of each independent model to capture the evolution characteristics of solar flares, which is a much better performance than traditional statistical prediction models or any single machine-learning method. We also proposed two frameworks, namely F1_FFM and TSS_FFM, which optimize the F1 score and TSS score, respectively. The cross validation results show that they have their respective advantages in the F1 score and TSS score.

Improved and Interpretable Solar Flare Predictions With Spatial and Topological Features of the Polarity Inversion Line Masked Magnetograms

AbstractMany current research efforts undertake the solar flare classification task using the Space‐weather HMI Active Region Patch (SHARP) parameters as the predictors. The SHARP parameters are scalar quantities based on spatial average or integration of physical quantities derived from the vector magnetic field, which loses information of the two‐dimensional spatial distribution of the field and related quantities. In this paper, we construct two new sets of spatial features to expand the feature set used for the flare classification task. The first set uses the idea of topological data analysis to summarize the geometric information of the distributions of various SHARP quantities across active regions. The second set utilizes tools coming from spatial statistics to analyze the vertical magnetic field component Br and summarize its spatial variations and clustering patterns. All features are constructed within regions near the polarity inversion lines (PILs) and classification performances using the new features are compared against those using SHARP parameters (also along the PIL). We found that using the new features can improve the skill scores of the flare classification model and new features tend to have higher feature importance, especially the spatial statistics features. This potentially suggests that even using a single magnetic field component, Br, instead of all SHARP parameters, one can still derive strongly predictive features for flare classification.

Proteomic Approaches to Defining Remission and the Risk of Relapse in Rheumatoid Arthritis.

ObjectivesPatients with Rheumatoid Arthritis (RA) are increasingly achieving stable disease remission, yet the mechanisms that govern ongoing clinical disease and subsequent risk of future flare are not well understood. We sought to identify serum proteomic alterations that dictate clinically important features of stable RA, and couple broad-based proteomics with machine learning to predict future flare.MethodsWe studied baseline serum samples from a cohort of stable RA patients (RETRO, n = 130) in clinical remission (DAS28<2.6) and quantified 1307 serum proteins using the SOMAscan platform. Unsupervised hierarchical clustering and supervised classification were applied to identify proteomic-driven clusters and model biomarkers that were associated with future disease flare after 12 months of follow-up and RA medication withdrawal. Network analysis was used to define pathways that were enriched in proteomic datasets.ResultsWe defined 4 proteomic clusters, with one cluster (Cluster 4) displaying a lower mean DAS28 score (p = 0.03), with DAS28 associating with humoral immune responses and complement activation. Clustering did not clearly predict future risk of flare, however an XGboost machine learning algorithm classified patients who relapsed with an AUC (area under the receiver operating characteristic curve) of 0.80 using only baseline serum proteomics.ConclusionsThe serum proteome provides a rich dataset to understand stable RA and its clinical heterogeneity. Combining proteomics and machine learning may enable prediction of future RA disease flare in patients with RA who aim to withdrawal therapy.

Class imbalance problem in short-term solar flare prediction

Using data-driven algorithms to accurately forecast solar flares requires reliable data sets. The solar flare dataset is composed of many non-flaring samples with a small percentage of flaring samples. This is called the class imbalance problem in data mining tasks. The prediction model is sensitive to most classes of the original data set during training. Therefore, the class imbalance problem for building up the flare prediction model from observational data should be systematically discussed. Aiming at the problem of class imbalance, three strategies are proposed corresponding to the data set, loss function, and training process: Type I resamples the training samples, including oversampling for the minority class, undersampling, or mixed sampling for the majority class. Type II usually changes the decision-making boundary, assigning the majority and minority categories of prediction loss to different weights. Type III assigns different weights to the training samples, the majority categories are assigned smaller weights, and the minority categories are assigned larger weights to improve the training process of the prediction model. The main work of this paper compares these imbalance processing methods when building a flare prediction model and tries to find the optimal strategy. Our results show that among these strategies, the performance of oversampling and sample weighting is better than other strategies in most parameters, and the generality of resampling and changing the decision boundary is better.

Serum Calprotectin a Potential Biomarker in Juvenile Idiopathic Arthritis: A Meta-Analysis.

Juvenile idiopathic arthritis (JIA) is the most common inflammatory chronic disease affecting children and adolescents. Today, there are no specific biomarkers of inflammation. Therefore, it is important to identify new markers as predictors of disease activity. Recently, some researchers have directed their interest toward a protein, calprotectin (CLP), as a potential biomarker. The primary objective of our systematic review and meta-analysis was to analyze the possible role of CLP in JIA. Method: A literature search was conducted using PubMed, EMBASE, Scopus, Science Direct on 10 August 2021. The selection of studies was made using the PRISMA 2020 guidelines. Cohen’s d with 95% CI and p-value were used as a measure of effect size. The random effects model was used to account for different sources of variation among studies. Heterogeneity was assessed using Q statistics and I2. The publication bias was analyzed and represented by a funnel plot, and funnel plot symmetry was assessed with Egger’s test. Results: Our results at follow-up showed a statistically significant difference between patients with active disease compared to patients with inactive disease: 0.39 (0.16; 0.62), p = 0.001; without statistical heterogeneity. Another important aspect that emerged were the differences between the systemic disease form and any form of inactive disease showing a different concentration of calprotectin: 0.74 (0.40; 1.08), p < 0.001; without statistical heterogeneity. On the other hand, meta-regression analyses performed on gender, age, duration of disease, percentage of patients with ANA+ or RF+, medium value of ESR or CRP were not statistically significant. A statistically significant difference in serum calprotectin concentration between patients with JIA and healthy controls were observed. In fact, it presented lower values in the control group. Conclusions: The use of serum CLP could represent, in the future, a useful tool in JIA in order to stratify disease activity more accurately and may aid a more tailored approach to drug of choice in children with JIA. Further studies are needed to evaluate CLP as a predictor of flare in combination with other potential biomarkers of subclinical disease activity.

S716 Clinical Predictors of Disease Flares in Pregnant People With Inflammatory Bowel Disease

Introduction: Pregnancy in people with inflammatory bowel disease (IBD) is associated with a higher risk of adverse outcomes compared to the general population. Identification of clinical factors which predict flares could be used to guide treatment decisions. Our aim was to identify clinical predictors of disease flares in pregnant people with IBD. Methods: A retrospective cohort study of all deliveries from 1/1/2011 to 1/1/2020 at an academic tertiary care medical center in a major metropolitan area in the United States. Pregnant people were identified using ICD-9 codes for IBD and CPT codes for deliveries. Charts were reviewed for baseline demographics, obstetric history, disease subtype (Crohn’s disease [CD] or ulcerative colitis [UC]), medication use, disease course and adverse outcomes during pregnancy. Results: 42,000 deliveries were identified of which 122 were in those with IBD of whom 33% (n = 40) flared during pregnancy. UC diagnosis was associated with higher risk of pregnancy flare compared to CD (55% vs 22.4% p=0.001). Elevated stool calprotectin was associated with higher risk of pregnancy flare (564.4 vs 37.6 p=0.007). Patients on thiopurines and combination therapy with a biologic had fewer flares than those on no medications, 5-ASAs, biologics alone or steroids (p< 0.05). Pregnant people in endoscopic or clinical remission prior to pregnancy had significantly fewer disease flares compared to those who did not (p=0.0001). Pregnant people who flared had significantly higher rates of preterm premature rupture of membranes (PPROM 10.5% vs 1.3% p=0.04) and small for gestational age offspring (12.8% vs 2.6% p=0.04) compared to those who did not experience a peripartum flare (Table 1). Rate of other adverse obstetric and neonatal outcomes did not differ between groups (Table 1). A logistic regression model utilizing clinical, endoscopic remission, and calprotectin resulted in 85.7% accuracy for prediction of flares. Conclusion: In our population of pregnant people with IBD, disease flares during pregnancy were associated with higher rates preterm premature rupture of membranes and small for gestational age offspring. Clinical and endoscopic remission along with calprotectin data may be sufficient to predict disease flares during pregnancy and assist in counseling and treating this group of patients to maximize optimal pregnancy outcomes.Table 1.: Patient Characteristics and Outcomes.

DeepSun: machine-learning-as-a-service for solar flare prediction

Solar flare prediction plays an important role in understanding and forecasting space weather. The main goal of the Helioseismic and Magnetic Imager (HMI), one of the instruments on NASA’s Solar Dynamics Observatory, is to study the origin of solar variability and characterize the Sun’s magnetic activity. HMI provides continuous full-disk observations of the solar vector magnetic field with high cadence data that lead to reliable predictive capability; yet, solar flare prediction effort utilizing these data is still limited. In this paper, we present a machine-learning-as-a-service (MLaaS) framework, called DeepSun, for predicting solar flares on the web based on HMI’s data products. Specifically, we construct training data by utilizing the physical parameters provided by the Space-weather HMI Active Region Patch (SHARP) and categorize solar flares into four classes, namely B, C, M and X, according to the X-ray flare catalogs available at the National Centers for Environmental Information (NCEI). Thus, the solar flare prediction problem at hand is essentially a multi-class (i.e., four-class) classification problem. The DeepSun system employs several machine learning algorithms to tackle this multi-class prediction problem and provides an application programming interface (API) for remote programming users. To our knowledge, DeepSun is the first MLaaS tool capable of predicting solar flares through the internet.

Hybrid deep convolutional neural network with one-versus-one approach for solar flare prediction

ABSTRACT We propose a novel hybrid Convolutional Neural Network (CNN) model with one-versus-one approach to forecast solar flare occurrence with the outputs of four classes (No-flare, C, M, and X) within 24 h. We train and test our model using the same data sets as in Zheng, Li &amp; Wang, and then compare our results with previous models using the true skill statistic (TSS) as primary metric. The main results are as follows. (1) This is the first time that the CNN model in conjunction with one-versus-one approach is used in solar physics to make multiclass flare prediction. (2) In the four-class flare prediction, our model achieves quite high mean scores of TSS = 0.703, 0.489, 0.432, and 0.436 for No-flare, C, M, and X class, respectively, which are much better than or comparable to those of previous studies. In addition, our model obtains TSS scores of 0.703 ± 0.070 for ≥C-class and 0.739 ± 0.109 for ≥M-class predictions. (3) This is the first attempt to open the black-box CNN model to study the visualization of feature maps for interpreting the prediction model. Furthermore, the visualization results indicate that our model pays attention to the regions with strong gradient, strong intensity, high total intensity, and large range of the intensity in high-level feature maps. The median gradient and intensity, the total intensity, and the range of the intensity for high-level feature maps increase approximately with the increase of flare level.

Shape-based Feature Engineering for Solar Flare Prediction

Solar flares are caused by magnetic eruptions in active regions (ARs) on the surface of the sun. These events can have significant impacts on human activity, many of which can be mitigated with enough advance warning from good forecasts. To date, machine learning-based flare-prediction methods have employed physics-based attributes of the AR images as features; more recently, there has been some work that uses features deduced automatically by deep learning methods (such as convolutional neural networks). We describe a suite of novel shape-based features extracted from magnetogram images of the Sun using the tools of computational topology and computational geometry. We evaluate these features in the context of a multi-layer perceptron (MLP) neural network and compare their performance against the traditional physics-based attributes. We show that these abstract shape-based features outperform the features chosen by the human experts, and that a combination of the two feature sets improves the forecasting capability even further.

Immunosuppressive therapy withdrawal after remission achievement in patients with lupus nephritis.

Whether immunosuppressive therapy may be safely withdrawn in lupus nephritis (LN) is still unclear. We assessed rate and predictors of flare after IS withdrawal in patients with LN in remission. Patients with biopsy-proven LN treated with immunosuppressants (IS) between 1980 and 2020 were considered. Remission was defined as normal serum creatinine, proteinuria <0.5 g/24 h, inactive urine sediment, and no extra-renal SLE activity on stable immunosuppressive and/or antimalarial therapy and/or prednisone ≤5mg/day. IS discontinuation was defined as the complete withdrawal of immunosuppressive therapy, flares according to SLEDAI Flare Index. Predictors of flare were analysed by multivariate logistic regression analysis. Among 513 SLE patients included in our database, 270 had LN. Of them, 238 underwent renal biopsy and were treated with IS. Eighty-three patients (34.8%) discontinued IS, 46 (30) months after remission achievement. During a mean (s.d.) follow-up of 116.5 (78) months, 19 patients (22.9%) developed a flare (8/19 renal) and were re-treated; 14/19 (73.7%) re-achieved remission after restarting therapy. Patients treated with IS therapy for at least 3 years after remission achievement had the lowest risk of relapse (OR 0.284, 95% CI: 0.093, 0.867; P = 0.023). At multivariate analysis, antimalarial maintenance therapy (OR 0.194, 95% CI: 0.038, 0.978; P = 0.047), age at IS discontinuation (OR 0.93, 95% CI: 0.868, 0.997; P = 0.040), remission duration >3 years before IS discontinuation (OR 0.231, 95% CI: 0.058, 0.920; P = 0.038) were protective against disease flares. Withdrawal of IS is feasible in LN patients in remission for at least 3 years and on antimalarial therapy. Patients who experience flares can re-achieve remission with an appropriate treatment.

Prediction of disease flare by biomarkers after discontinuing biologics in patients with rheumatoid arthritis achieving stringent remission

To elucidate the disease-flare process in rheumatoid arthritis (RA) after discontinuing biological disease-modifying antirheumatic drugs (bDMARDs), we first focused on RA-flare prediction after achieving stringent remission criteria. Patients with RA who maintained a simplified disease activity index ≤ 3.3 for ≥ 3 months during November 2014–January 2018 in our medical centre in Tokyo, Japan, were eligible. The primary endpoint was flare (disease activity score 28—erythrocyte sedimentation rate ≥ 3.2 with increase from baseline > 0.6) within 2 years after bDMARD discontinuation. Comprehensive clinical assessments, ultrasonographic evaluation of 40 joints, and blood sampling for 12 biomarkers were performed every 2–3 months for 2 years unless patients experienced flare. Flare-positive and flare-negative patients were compared using univariate and Kaplan–Meier analyses. Thirty-six patients (80.6% female, median disease duration, 5.2 years; median treatment period with discontinued bDMARD, 2 years; median remission duration, 18 months) were enrolled. Twenty patients (55.6%) experienced RA flare 43–651 (median, 115) days after the first skipped date of bDMARDs. Two patients who withdrew without disease flare were excluded from the comparison. Clinical and ultrasonographic evaluations did not show significant between-group differences; Kaplan–Meier analysis showed that higher baseline soluble tumour necrosis factor receptor 1 (sTNFR1) concentration impacted subsequent disease flare (p = 0.0041); higher baseline interleukin (IL)-2 concentration was exclusively beneficial to patients with lower sTNFR1 (p = 0.0058), resulting in remission maintenance in 83.3% of patients with lower sTNFR1 and higher IL-2. We demonstrated the usefulness of combined biomarker evaluation for predicting sustained remission after bDMARD discontinuation in RA.

Machine Learning for Prediction and Risk Stratification of Lupus Nephritis Renal Flare

Background: Renal flare of lupus nephritis (LN) is strongly associated with poor kidney outcomes, and predicting renal flare and stratifying its risk are important for clinical decision-making and individualized management to reduce LN flare. Methods: We randomly divided 1,694 patients with biopsy-proven LN, who had achieved remission after treatment, into a derivation cohort (n = 1,186) and an internal validation cohort (n = 508), at a ratio of 7:3. The risk of renal flare 5 years after remission was predicted using an eXtreme Gradient Boosting (XGBoost) method model, developed from 59 variables, including demographic, clinical, immunological, pathological, and therapeutic characteristics. A simplified risk score prediction model (SRSPM) was developed from important variables selected by XGBoost model using stepwise Cox regression for practical convenience. Results: The 5-year relapse rates were 39.5% and 38.2% in the derivation and internal validation cohorts, respectively. Both the XGBoost model and the SRSPM had good predictive performance, with a C-index of 0.819 (95% confidence interval [CI]: 0.774–0.857) and 0.746 (95% CI: 0.697–0.795), respectively, in the validation cohort. The SRSPM comprised 6 variables, including partial remission and endocapillary hypercellularity at baseline, age, serum Alb, anti-dsDNA, and serum complement C3 at the point of remission. Using Kaplan-Meier analysis, the SRSPM identified significant risk stratification for renal flares (p < 0.001). Conclusions: Renal flare of LN can be readily predicted using the XGBoost model and the SRSPM, and the SRSPM can also stratify flare risk. Both models are useful for clinical decision-making and individualized management in LN.

Which patients with systemic lupus erythematosus in remission can withdraw low dose steroids? Results from a single inception cohort study.

A progressive tapering until withdrawal of glucocorticoids (GC) is considered one of the main goals of Systemic Lupus Erythematosus (SLE) management. However, which patient may be a candidate for safe GC withdrawal has not been determined yet. This study aimed to evaluate the rate of low-dose GC withdrawal in SLE patients in remission and to identify predictors of flares. Eligible patients were SLE patients in prolonged clinical remission defined by a cSLEDAI = 0 for at least 2 years and on a stable SLE treatment (including daily 5 mg prednisone). Flares were defined by SELENA-SLEDAI Flare Index. Predictors of flares after GC withdrawal were analyzed by Cox regression. We selected 56 patients in whom a GC withdrawal was attempted. 98 patients were in the prednisone maintenance group. The proportion of patients experiencing a flare was not significantly lower in the maintenance group than in the withdrawal group (p = 0.81). However, among the withdrawal group, the rate of flares was significantly higher in serologically active clinically quiescent (SACQ) patients (p < 0,0001). At Cox regression analysis, duration of hydroxychloroquine (HCQ) therapy and ≥5 year remission at withdrawal were protective factors, while a SACQ disease and history of lupus nephritis increased the risk of disease flare. GC withdrawal is an achievable target in SLE and may be attempted in patients in complete remission.However, it might underline a caution in patients with SACQ disease who may be at greater risk forflare when GCare discontinued. HCQ therapy and durable remission can significantly reduce the risk.

Operational solar flare prediction model using Deep Flare Net

We developed an operational solar flare prediction model using deep neural networks, named Deep Flare Net (DeFN). DeFN can issue probabilistic forecasts of solar flares in two categories, such as ≥ M-class and < M-class events or ≥ C-class and < C-class events, occurring in the next 24 h after observations and the maximum class of flares occurring in the next 24 h. DeFN is set to run every 6 h and has been operated since January 2019. The input database of solar observation images taken by the Solar Dynamic Observatory (SDO) is downloaded from the data archive operated by the Joint Science Operations Center (JSOC) of Stanford University. Active regions are automatically detected from magnetograms, and 79 features are extracted from each region nearly in real time using multiwavelength observation data. Flare labels are attached to the feature database, and then, the database is standardized and input into DeFN for prediction. DeFN was pretrained using the datasets obtained from 2010 to 2015. The model was evaluated with the skill score of the true skill statistics (TSS) and achieved predictions with TSS = 0.80 for ≥ M-class flares and TSS = 0.63 for ≥ C-class flares. For comparison, we evaluated the operationally forecast results from January 2019 to June 2020. We found that operational DeFN forecasts achieved TSS = 0.70 (0.84) for ≥ C-class flares with the probability threshold of 50 (40)%, although there were very few M-class flares during this period and we should continue monitoring the results for a longer time. Here, we adopted a chronological split to divide the database into two for training and testing. The chronological split appears suitable for evaluating operational models. Furthermore, we proposed the use of time-series cross-validation. The procedure achieved TSS = 0.70 for ≥ M-class flares and 0.59 for ≥ C-class flares using the datasets obtained from 2010 to 2017. Finally, we discuss the standard evaluation methods for operational forecasting models, such as the preparation of observation, training, and testing datasets, and the selection of verification metrics.

Advanced machine learning for predicting individual risk of flares in rheumatoid arthritis patients tapering biologic drugs

BackgroundBiological disease-modifying anti-rheumatic drugs (bDMARDs) can be tapered in some rheumatoid arthritis (RA) patients in sustained remission. The purpose of this study was to assess the feasibility of building a model to estimate the individual flare probability in RA patients tapering bDMARDs using machine learning methods.MethodsLongitudinal clinical data of RA patients on bDMARDs from a randomized controlled trial of treatment withdrawal (RETRO) were used to build a predictive model to estimate the probability of a flare. Four basic machine learning models were trained, and their predictions were additionally combined to train an ensemble learning method, a stacking meta-classifier model to predict the individual flare probability within 14 weeks after each visit. Prediction performance was estimated using nested cross-validation as the area under the receiver operating curve (AUROC). Predictor importance was estimated using the permutation importance approach.ResultsData of 135 visits from 41 patients were included. A model selection approach based on nested cross-validation was implemented to find the most suitable modeling formalism for the flare prediction task as well as the optimal model hyper-parameters. Moreover, an approach based on stacking different classifiers was successfully applied to create a powerful and flexible prediction model with the final measured AUROC of 0.81 (95%CI 0.73–0.89). The percent dose change of bDMARDs, clinical disease activity (DAS-28 ESR), disease duration, and inflammatory markers were the most important predictors of a flare.ConclusionMachine learning methods were deemed feasible to predict flares after tapering bDMARDs in RA patients in sustained remission.

The automated prediction of solar flares from SDO images using deep learning

In the last few years, there has been growing interest in near-real-time solar data processing, especially for space weather applications. This is due to space weather impacts on both space-borne and ground-based systems, and industries, which subsequently impacts our lives. In the current study, the deep learning approach is used to establish an automated hybrid computer system for a short-term forecast; it is achieved by using the complexity level of the sunspot group on SDO/HMI Intensitygram images. Furthermore, this suggested system can generate the forecast for solar flare occurrences within the following 24 h. The input data for the proposed system are SDO/HMI full-disk Intensitygram images and SDO/HMI full-disk magnetogram images. System outputs are the “Flare or Non-Flare” of daily flare occurrences (C, M, and X classes). This system integrates an image processing system to automatically detect sunspot groups on SDO/HMI Intensitygram images using active-region data extracted from SDO/HMI magnetogram images (presented by Colak and Qahwaji, 2008) and deep learning to generate these forecasts. Our deep learning-based system is designed to analyze sunspot groups on the solar disk to predict whether this sunspot group is capable of releasing a significant flare or not. Our system introduced in this work is called ASAP_Deep. The deep learning model used in our system is based on the integration of the Convolutional Neural Network (CNN) and Softmax classifier to extract special features from the sunspot group images detected from SDO/HMI (Intensitygram and magnetogram) images. Furthermore, a CNN training scheme based on the integration of a back-propagation algorithm and a mini-batch AdaGrad optimization method is suggested for weight updates and to modify learning rates, respectively. The images of the sunspot regions are cropped automatically by the imaging system and processed using deep learning rules to provide near real-time predictions. The major results of this study are as follows. Firstly, the ASAP_Deep system builds on the ASAP system introduced in Colak and Qahwaji (2009) but improves the system with an updated deep learning-based prediction capability. Secondly, we successfully apply CNN to the sunspot group image without any pre-processing or feature extraction. Thirdly, our system results are considerably better, especially for the false alarm ratio (FAR); this reduces the losses resulting from the protection measures applied by companies. Also, the proposed system achieves a relatively high scores for True Skill Statistics (TSS) and Heidke Skill Score (HSS).

Anti-myeloperoxidase and proteinase 3 antibodies for nephritis flare prediction in anti-neutrophil cytoplasmic antibody-associated vasculitis.

The value of myeloperoxidase (MPO) and proteinase 3 (PR3) antibody titres in the assessment of renal disease activity and flare prediction in patients with anti-neutrophil cytoplasmic antibody-associated vasculitis (AAV) is not well known. We performed a retrospective study including 113 AVV patients with renal biopsy-proven pauci-immune necrotizing glomerulonephritis from seven Spanish hospitals. The main inclusion criteria were assessment of MPO antibodies using multiplex flow immunoassay and PR3 antibody measurements using immunoassay chemiluminescence with an identical range of values for all participating centres. Serum MPO antibodies 3 ± 1.2 months before relapse were higher in patients who relapsed [19.2 ± 12.2 versus 3.2 ± 5.1 antibody index (AI); P < 0.001]. The discrimination value of MPO antibodies 3 months before renal relapse had an area under the receiver operating characteristics curve (AUC) of 0.82 [95% confidence interval (CI) 0.73-0.92; P < 0.001]. ΔMPO antibodies (change in antibodies titration 6 months before relapse) were higher in patients who relapsed (8.3 ± 12 versus 0.9 ± 3.1 AI; P = 0.001). The discrimination value of ΔMPO had an AUC of 0.76 (95% CI 0.63-0.88; P < 0.001). The positive predictive value of renal relapse in PR3 patients is 100% and the negative predictive value of renal relapse in patients with PR3-positive titres is 57.1%. Serum PR3 antibodies were higher in patients who relapsed 2.8 ± 1.4 months before relapse (58.6 ± 24.6 versus 2.0 ± 0.6 AI; P < 0.001). MPO level monitoring using multiplex flow immunoassay and PR3 measurements using immunoassay chemiluminescence are useful and sensitive tools for the prediction of renal relapse in the follow-up of AAV patients with renal disease and relevant surrogate markers of renal disease activity.