Prediction of lupus nephritis in patients with systemic lupus erythematosus using artificial neural networks

Gene-Based Predictive Modelling for Enhanced Detection of Systemic Lupus Erythematosus Using CNN-Based DL Algorithm

Systemic Lupus Erythematosus (SLE) is a multifaceted autoimmune disease that presents with a diverse array of clinical signs and unpredictable disease progression. Conventional diagnostic methods frequently fall short in terms of sensitivity and specificity, which can result in delayed diagnosis and less-than-optimal management. In this study, we introduce a novel approach for improving the identification of SLE through the use of gene-based predictive modelling and Stacked deep learning classifiers. The study proposes a new method for diagnosing SLE using Stacked Deep Learning Classifiers (SDLC) trained on Gene Expression Omnibus (GEO) database data. By combining transcriptomic data from GEO with clinical features and laboratory results, the SDLC model achieves a remarkable accuracy value of 0.996, outperforming traditional methods. Individual models within the SDLC, such as SBi-LSTM and ACNN, achieved accuracies of 92% and 95%, respectively. The SDLC's ensemble learning approach allows for identifying complex patterns in multi-modal data, enhancing accuracy in diagnosing SLE. This study emphasises the potential of deep learning methods, in conjunction with open repositories like GEO, to advance the diagnosis and management of SLE. Overall, this research shows strong performance and potential for improving precision medicine in managing SLE.

Artificial intelligence and high-dimensional technologies in the theragnosis of systemic lupus erythematosus

Systemic lupus erythematosus is a complex, systemic autoimmune disease characterised by immune dysregulation. Pathogenesis is multifactorial, contributing to clinical heterogeneity and posing challenges for diagnosis and treatment. Although strides in treatment options have been made in the past 15 years, with the US Food and Drug Administration approval of belimumab in 2011, there are still many patients who have inadequate responses to therapy. A better understanding of underlying disease mechanisms with a holistic and multiparametric approach is required to improve clinical assessment and treatment. This Review discusses the evolution of genomics, epigenomics, transcriptomics, and proteomics in the study of systemic lupus erythematosus and ways to amalgamate these silos of data with a systems-based approach while also discussing ways to strengthen the overall process. These mechanistic insights will facilitate the discovery of functionally relevant biomarkers to guide rational therapeutic selection to improve patient outcomes.

Current state and completeness of reporting clinical prediction models using machine learning in systemic lupus erythematosus: A systematic review

OBJECTIVE

We carried out a systematic review (SR) of adherence in diagnostic and prognostic applications of ML in SLE using the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) Statement.

METHODS

A SR employing five databases was conducted from its inception until December 2021. We identified articles that evaluated the utilization of ML for prognostic and/or diagnostic purposes. This SR was reported based on the PRISMA guidelines. The TRIPOD statement assessed adherence to reporting standards. Assessment for risk of bias was done using PROBAST tool.

RESULTS

We included 45 studies: 29 (64.4%) diagnostic and 16 (35.5%) prognostic prediction- model studies. Overall, articles adhered by between 17% and 67% (median 43%, IQR 37-49%) to TRIPOD items. Only few articles reported the model's predictive performance (2.3%, 95% CI 0.06-12.0), testing of interaction terms (2.3%, 95% CI 0.06-12.0), flow of participants (50%, 95% CI; 34.6-65.4), blinding of predictors (2.3%, 95% CI 0.06-12.0), handling of missing data (36.4%, 95% CI 22.4-52.2), and appropriate title (20.5%, 95% CI 9.8-35.3). Some items were almost completely reported: the source of data (88.6%, 95% CI 75.4-96.2), eligibility criteria (86.4%, 95% CI 76.2-96.5), and interpretation of findings (88.6%, 95% CI 75.4-96.2). In addition, most of model studies had high risk of bias.

CONCLUSIONS

The reporting adherence of ML-based model developed for SLE, is currently inadequate. Several items deemed crucial for transparent reporting were not fully reported in studies on ML-based prediction models.

REVIEW REGISTRATION

PROSPERO ID# CRD42021284881. (Amended to limit the scope).

Artificial intelligence in glomerular diseases

In this narrative review, we focus on the application of artificial intelligence in the clinical history of patients with glomerular disease, digital pathology in kidney biopsy, renal ultrasonography imaging, and prediction of chronic kidney disease (CKD). With the development of natural language processing, the clinical history of a patient can be used to identify a computable phenotype. In kidney pathology, digital imaging has adopted innovative deep learning algorithms (DLAs) that can improve the predictive capability of the examined lesions. However, at this time, these applications can only be used in research because there is no recognized validation to replace the conventional diagnostic applications. Kidney ultrasonography, used in the clinical examination of patients, provides information about the progression of kidney damage. Machine learning algorithms (MLAs) with promising results for the early detection of CKD have been proposed, but, still, they are not solid enough to be incorporated into the clinical practice. A few tools for glomerulonephritis, based on MLAs, are available in clinical practice. They can be downloaded on computers and cellular phones but can only be applied to uniracial cohorts of patients. To improve their performance, it is necessary to organize large consortia with multiracial cohorts. Finally, in many studies MLA development has been carried out using retrospective cohorts. The performance of the models might differ in retrospective cohorts compared to real-world data. Therefore, the models should be validated in prospective external large cohorts.

Development and Verify of Survival Analysis Models for Chinese Patients With Systemic Lupus Erythematosus

Background

The aim of this study is to develop survival analysis models of hospitalized systemic lupus erythematosus (h-SLE) patients in Jiangsu province using data mining techniques to predict patient survival outcomes and survival status.

Methods

In this study, based on 1999–2009 survival data of 2453 hospitalized SLE (h-SLE) patients in Jiangsu Province, we not only used the Cox proportional hazards model to analyze patients’ survival factors, but also used neural network models to predict survival outcomes. We used semi-supervised learning to label the censored data and introduced cost-sensitivity to achieve data augmentation, addressing category imbalance and pseudo label credibility. In addition, the risk score model was developed by logistic regression.

Results

The overall accuracy of the survival outcome prediction model exceeded 0.7, and the sensitivity was close to 0.8, and through the comparative analysis of multiple indicators, our model outperformed traditional classifiers. The developed survival risk assessment model based on logistic regression found that there was a clear threshold, i.e., a survival threshold indicating the survival risk of patients, and cardiopulmonary and neuropsychiatric involvement, abnormal blood urea nitrogen levels and alanine aminotransferase level had the greatest impact on patient survival time. In addition, the study developed a graphical user interface (GUI) integrating survival analysis models to assist physicians in diagnosis and treatment.

Conclusions

The proposed survival analysis scheme identifies disease-related pathogenic and prognosis factors, and has the potential to improve the effectiveness of clinical interventions.

Prediction of diagnosis results of rheumatoid arthritis patients based on autoantibodies and cost-sensitive neural network

OBJECTIVES

To analyze and evaluate the effectiveness of the detection of single autoantibody and combined autoantibodies in patients with rheumatoid arthritis (RA) and related autoimmune diseases and establish a machine learning model to predict the disease of RA.

METHODS

A total of 309 patients with joint pain as the first symptom were retrieved from the database. The effectiveness of single and combined antibodies tests was analyzed and evaluated in patients with RA, a cost-sensitive neural network (CSNN) model was used to integrate multiple autoantibodies and patient symptoms to predict the diagnosis of RA, and the ROC curve was used to analyze the diagnosis performance and calculate the optimal cutoff value.

RESULTS

There are differences in the seropositive rate of autoimmune diseases, the sensitivity and specificity of single or multiple autoantibody tests were insufficient, and anti-CCP performed best in RA diagnosis and had high diagnostic value. The cost-sensitive neural network prediction model had a sensitivity of up to 0.90 and specificity of up to 0.86, which was better than a single antibody and combined multiple antibody detection.

CONCLUSION

In-depth analysis of autoantibodies and reliable early diagnosis based on the neural network could guide specialized physicians to develop different treatment plans to prevent deterioration and enable early treatment with antirheumatic drugs for remission. Key Points • There are differences in the seropositive rate of autoimmune diseases. • This is the first study to use a cost-sensitive neural network model to diagnose RA disease in patients. • The diagnosis effect of the cost-sensitive neural network model is better than a single antibody and combined multiple antibody detection.

The pathogenesis of systemic lupus erythematosus: Harnessing big data to understand the molecular basis of lupus

Systemic lupus erythematosus (SLE) is a chronic, systemic autoimmune disease that causes damage to multiple organ systems. Despite decades of research and available murine models that capture some aspects of the human disease, new treatments for SLE lag behind other autoimmune diseases such as Rheumatoid Arthritis and Crohn's disease. Big data genomic assays have transformed our understanding of SLE by providing important insights into the molecular heterogeneity of this multigenic disease. Gene wide association studies have demonstrated more than 100 risk loci, supporting a model of multiple genetic hits increasing SLE risk in a non-linear fashion, and providing evidence of ancestral diversity in susceptibility loci. Epigenetic studies to determine the role of methylation, acetylation and non-coding RNAs have provided new understanding of the modulation of gene expression in SLE patients and identified new drug targets and biomarkers for SLE. Gene expression profiling has led to a greater understanding of the role of myeloid cells in the pathogenesis of SLE, confirmed roles for T and B cells in SLE, promoted clinical trials based on the prominent interferon signature found in SLE patients, and identified candidate biomarkers and cellular signatures to further drug development and drug repurposing. Gene expression studies are advancing our understanding of the underlying molecular heterogeneity in SLE and providing hope that patient stratification will expedite new therapies based on personal molecular signatures. Although big data analyses present unique interpretation challenges, both computationally and biologically, advances in machine learning applications may facilitate the ability to predict changes in SLE disease activity and optimize therapeutic strategies.

Prediction of chronic damage in systemic lupus erythematosus by using machine-learning models

Objective

The increased survival in Systemic Lupus Erythematosus (SLE) patients implies the development of chronic damage, occurring in up to 50% of cases. Its prevention is a major goal in the SLE management. We aimed at predicting chronic damage in a large monocentric SLE cohort by using neural networks.

Methods

We enrolled 413 SLE patients (M/F 30/383; mean age ± SD 46.3±11.9 years; mean disease duration ± SD 174.6 ± 112.1 months). Chronic damage was assessed by the SLICC/ACR Damage Index (SDI). We applied Recurrent Neural Networks (RNNs) as a machine-learning model to predict the risk of chronic damage. The clinical data sequences registered for each patient during the follow-up were used for building and testing the RNNs.

Results

At the first visit in the Lupus Clinic, 35.8% of patients had an SDI≥1. For the RNN model, two groups of patients were analyzed: patients with SDI = 0 at the baseline, developing damage during the follow-up (N = 38), and patients without damage (SDI = 0). We created a mathematical model with an AUC value of 0.77, able to predict damage development. A threshold value of 0.35 (sensitivity 0.74, specificity 0.76) seemed able to identify patients at risk to develop damage.

Conclusion

We applied RNNs to identify a prediction model for SLE chronic damage. The use of the longitudinal data from the Sapienza Lupus Cohort, including laboratory and clinical items, resulted able to construct a mathematical model, potentially identifying patients at risk to develop damage.

Application of Various Statistical Models to Explore Gene-Gene Interactions in Folate, Xenobiotic, Toll-Like Receptor and STAT4 Pathways that Modulate Susceptibility to Systemic Lupus Erythematosus

INTRODUCTION

In view of our previous studies showing an independent association of genetic polymorphisms in folate, xenobiotic, and toll-like receptor (TLR) pathways with the risk for systemic lupus erythematosus (SLE), we have developed three statistical models to delineate complex gene-gene interactions between folate, xenobiotic, TLR, and signal transducer and activator of transcription 4 (STAT4) signaling pathways in association with the molecular pathophysiology of SLE.

METHODS

We developed additive, multifactor dimensionality reduction (MDR), and artificial neural network (ANN) models.

RESULTS

The additive model, although the simplest, suggested a moderate predictability of 30 polymorphisms of these four pathways (area under the curve [AUC] 0.66). MDR analysis revealed significant gene-gene interactions among glutathione-S-transferase (GST)T1 and STAT4 (rs3821236 and rs7574865) polymorphisms, which account for moderate predictability of SLE. The MDR model for specific auto-antibodies revealed the importance of gene-gene interactions among cytochrome P450, family1, subfamily A, polypeptide 1 (CYP1A1) m1, catechol-O-methyltransferase (COMT) H108L, solute carrier family 19 (folate transporter), member 1 (SLC19A1) G80A, estrogen receptor 1 (ESR1), TLR5, 5-methyltetrahydrofolate-homocysteine methyltransferase reductase (MTRR), thymidylate synthase (TYMS). and STAT4 polymorphisms. The ANN model for disease prediction showed reasonably good predictability of SLE risk with 30 polymorphisms (AUC 0.76). These polymorphisms contribute towards the production of SSB and anti-dsDNA antibodies to the extent of 48 and 40%, respectively, while their contribution for the production of antiRNP, SSA, and anti-cardiolipin antibodies varies between 20 and 30%.

CONCLUSION

The current study highlighted the importance of genetic polymorphisms in folate, xenobiotic, TLR, and STAT4 signaling pathways as moderate predictors of SLE risk and delineates the molecular pathophysiology associated with these single nucleotide polymorphisms (SNPs) by demonstrating their association with specific auto-antibody production.

Development of Biomarker Models to Predict Outcomes in Lupus Nephritis

OBJECTIVE

The American College of Rheumatology guidelines for the treatment of lupus nephritis recommend change in induction therapy when response to therapy has not occurred within 6 months. Response is not defined, and renal fibrosis can occur while waiting for this end point. Therefore, a decision support tool to better define response is needed to guide clinicians when starting patients on therapy. This study was undertaken to identify biomarker models with sufficient predictive power to develop such a tool.

METHODS

Urine samples from 140 patients with biopsy-proven lupus nephritis who had not yet started induction therapy were analyzed for a panel of urinary biomarkers. Univariate receiver operating characteristic (ROC) curves were generated for each individual biomarker and compared to the ROC area under the curve values from machine learning models developed using random forest algorithms. Biomarker models of outcome developed with novel markers in addition to clinical markers were compared to those developed with traditional clinical markers alone.

RESULTS

Models developed with the combined traditional and novel biomarker panels demonstrated clinically meaningful predictive power. Markers most predictive of response were chemokines, cytokines, and markers of cellular damage.

CONCLUSION

This is the first study to demonstrate the power of low-abundance biomarker panels and machine learning algorithms for predicting lupus nephritis outcomes. This is a critical first step in research to develop clinically meaningful decision support tools.

Computer-assisted pattern recognition of autoantibody results

Immunoassay-based anti-nuclear antibody (ANA) screens are increasingly used in the initial evaluation of autoimmune disorders, but these tests offer no "pattern information" comparable to the information from indirect fluorescence assay-based screens. Thus, there is no indication of "next steps" when a positive result is obtained. To improve the utility of immunoassay-based ANA screening, we evaluated a new method that combines a multiplex immunoassay with a k nearest neighbor (kNN) algorithm for computer-assisted pattern recognition. We assembled a training set, consisting of 1,152 sera from patients with various rheumatic diseases and non-diseased patients. The clinical sensitivity and specificity of the multiplex method and algorithm were evaluated with a test set that consisted of 173 sera collected at a rheumatology clinic from patients diagnosed by using standard criteria, as well as 152 age- and sex-matched sera from presumably healthy individuals (sera collected at a blood bank). The test set was also evaluated with a HEp-2 cell-based enzyme-linked immunosorbent assay (ELISA). Both the ELISA and multiplex immunoassay results were positive for 94% of the systemic lupus erythematosus (SLE) patients. The kNN algorithm correctly proposed an SLE pattern for 84% of the antibody-positive SLE patients. For patients with no connective tissue disease, the multiplex method found fewer positive results than the ELISA screen, and no disease was proposed by the kNN algorithm for most of these patients. In conclusion, the automated algorithm could identify SLE patterns and may be useful in the identification of patients who would benefit from early referral to a specialist, as well as patients who do not require further evaluation.

Applying an artificial neural network to predict total body water in hemodialysis patients

BACKGROUND

Estimating total body water (TBW) is crucial in determining dry weight and dialytic dose for hemodialysis patients. Several anthropometric equations have been used to predict TBW, but a more accurate method is needed. We developed an artificial neural network (ANN) to predict TBW in hemodialysis patients.

METHODS

Demographic data, anthropometric measurements, and multifrequency bioelectrical impedance analysis (MF-BIA) were investigated in 54 patients. TBW measured by MF-BIA (TBW-BIA) was the reference. The predictive value of TBW based on ANN and five anthropometric equations (58% of actual body weight, Watson formula, Hume formula, Chertow formula, and Lee formula) was evaluated.

RESULTS

Predictive TBW values derived from anthropometric equations were significantly higher than TBW-BIA (31.341 +/- 6.033 liters). The only non-significant difference was between TBW-ANN (31.468 +/- 5.301 liters) and TBW-BIA (p = 0.639). ANN had the strongest Pearson's correlation coefficient (0.911) and smallest root mean square error (2.480); its peak centered most closely to zero with the shortest tails in an empirical cumulative distribution plot when compared with the other five equations.

CONCLUSION

ANN could surpass traditional anthropometric equations and serve as a feasible alternative method of TBW estimation for chronic hemodialysis patients.

Protein arrays and pattern recognition: new tools to assist in the identification and management of autoimmune disease

The occurrence of antibody patterns in connective tissue diseases has been recognized for thirty years, but the generation of multiple antibody results relied on time-consuming immunodiffusion or electrophoretic techniques. Today it is possible to study the antibody repertoire using rapid multi-analyte technologies, generally referred to as protein arrays. These arrays may use planar surfaces similar to DNA arrays, or use microspheres in suspension ("liquid arrays"). Also, many high quality autoantigens are now commercially available, including recombinant antigens. The vast amount of information that can be generated by measuring multiple antibodies for multiple patients has created demand for data processing. Software programs to aid physicians in reviewing multiple inputs as an aid to disease diagnosis and classification have been available for twenty years. Initial work used the "expert systems" approach; more recently pattern recognition has been widely evaluated because of the improvements in software programs and computational speed. The use of antibody data, generated in protein arrays, may assist in establishing diagnosis, in identifying potentially significant antibody patterns in advance of clinical symptoms, and in classifying patients based on expected disease progression.