Identification of biomarkers by machine learning classifiers to assist diagnose rheumatoid arthritis-associated interstitial lung disease

Clinically Relevant Biomarkers in Connective Tissue Disease-Associated Interstitial Lung Disease

Interstitial lung disease (ILD) complicates connective tissue disease (CTD) with variable incidence and is a leading cause of death in these patients. To improve CTD-ILD outcomes, early recognition and management of ILD is critical. Blood-based and radiologic biomarkers that assist in the diagnosis CTD-ILD have long been studied. Recent studies, including -omic investigations, have also begun to identify biomarkers that may help prognosticate such patients. This review provides an overview of clinically relevant biomarkers in patients with CTD-ILD, highlighting recent advances to assist in the diagnosis and prognostication of CTD-ILD.

Machine learning techniques for prediction in pregnancy complicated by autoimmune rheumatic diseases: Applications and challenges

Autoimmune rheumatic diseases are chronic conditions affecting multiple systems and often occurring in young women of childbearing age. The diseases and the physiological characteristics of pregnancy significantly impact maternal-fetal health and pregnancy outcomes. Currently, the integration of big data with healthcare has led to the increasing popularity of using machine learning (ML) to mine clinical data for studying pregnancy complications. In this review, we introduce the basics of ML and the recent advances and trends of ML in different prediction applications for common pregnancy complications by autoimmune rheumatic diseases. Finally, the challenges and future for enhancing the accuracy, reliability, and clinical applicability of ML in prediction have been discussed. This review will provide insights into the utilization of ML in identifying and assisting clinical decision-making for pregnancy complications, while also establishing a foundation for exploring comprehensive management strategies for pregnancy and enhancing maternal and child health.

Advancing precision rheumatology: applications of machine learning for rheumatoid arthritis management

Rheumatoid arthritis (RA) is an autoimmune disease causing progressive joint damage. Early diagnosis and treatment is critical, but remains challenging due to RA complexity and heterogeneity. Machine learning (ML) techniques may enhance RA management by identifying patterns within multidimensional biomedical data to improve classification, diagnosis, and treatment predictions. In this review, we summarize the applications of ML for RA management. Emerging studies or applications have developed diagnostic and predictive models for RA that utilize a variety of data modalities, including electronic health records, imaging, and multi-omics data. High-performance supervised learning models have demonstrated an Area Under the Curve (AUC) exceeding 0.85, which is used for identifying RA patients and predicting treatment responses. Unsupervised learning has revealed potential RA subtypes. Ongoing research is integrating multimodal data with deep learning to further improve performance. However, key challenges remain regarding model overfitting, generalizability, validation in clinical settings, and interpretability. Small sample sizes and lack of diverse population testing risks overestimating model performance. Prospective studies evaluating real-world clinical utility are lacking. Enhancing model interpretability is critical for clinician acceptance. In summary, while ML shows promise for transforming RA management through earlier diagnosis and optimized treatment, larger scale multisite data, prospective clinical validation of interpretable models, and testing across diverse populations is still needed. As these gaps are addressed, ML may pave the way towards precision medicine in RA.

Overview of Rheumatoid Arthritis-Associated Interstitial Lung Disease and Its Treatment

Interstitial lung disease (ILD) is a common pulmonary complication of rheumatoid arthritis (RA), causing significant morbidity and mortality. Optimal treatment for RA-ILD is not yet well defined. Reliable prognostic indicators are largely byproducts of prior ILD progression, including low or decreasing forced vital capacity and extensive or worsening fibrosis on imaging. In the absence of validated tools to predict treatment response, decisions about whether to initiate or augment treatment are instead based on clinical judgment. In general, treatment should be initiated in patients who are symptomatic, progressing, or at high risk of poor outcomes. Retrospective data suggest that mycophenolate mofetil, azathioprine, and rituximab are likely effective therapies for RA-ILD. Abatacept is also emerging as a potential first-line treatment option for patients with RA-ILD. Further, recent data demonstrate that immunosuppression may be beneficial even in patients with a usual interstitial pneumonia (UIP) pattern on imaging, suggesting that immunosuppression should be considered irrespective of imaging pattern. Recent randomized controlled trials have shown that antifibrotic medications, such as nintedanib and likely pirfenidone, slow forced vital capacity decline in RA-ILD. Consideration can be given to antifibrotic initiation in patients progressing despite immunosuppression, particularly in patients with a UIP pattern. Future research directions include developing tools to predict which patients will remain stable from patients who will progress, discriminating patients who will respond to treatment from nonresponders, and developing algorithms for starting immunosuppression, antifibrotics, or both as first-line therapies.

Research progress of KL-6 in respiratory system diseases

This article comprehensively elucidates the discovery of Krebs von den Lungen-6 (KL-6), its structural features, functional mechanisms, and the current research status in various respiratory system diseases. Discovered in 1985, KL-6 was initially considered a tumor marker, but its elevated levels in interstitial lung disease (ILD) led to its recognition as a relevant serum marker for ILD. KL-6 is primarily produced by type 2 alveolar epithelial cell regeneration. Over the past 30 years since the discovery of KL-6, the number of related research papers has steadily increased annually. Following the coronavirus disease 2019 (COVID-19) pandemic, there has been a sudden surge in relevant literature. Despite KL-6's potential as a biomarker, its value in the diagnosis, treatment, and prognosis varies across different respiratory diseases, including ILD, idiopathic pulmonary fibrosis (IPF), COVID-19, and lung cancer. Therefore, as an important serum biomarker in respiratory system diseases, the value of KL-6 still requires further investigation.

Machine learning application in autoimmune diseases: State of art and future prospectives

Autoimmune diseases are a group of disorders resulting from an alteration of immune tolerance, characterized by the formation of autoantibodies and the consequent development of heterogeneous clinical manifestations. Diagnosing autoimmune diseases is often complicated, and the available prognostic tools are limited. Machine learning allows us to analyze large amounts of data and carry out complex calculations quickly and with minimal effort. In this work, we examine the literature focusing on the use of machine learning in the field of the main systemic (systemic lupus erythematosus and rheumatoid arthritis) and organ-specific autoimmune diseases (type 1 diabetes mellitus, autoimmune thyroid, gastrointestinal, and skin diseases). From our analysis, interesting applications of machine learning emerged for developing algorithms useful in the early diagnosis of disease or prognostic models (risk of complications, therapeutic response). Subsequent studies and the creation of increasingly rich databases to be supplied to the algorithms will eventually guide the clinician in the diagnosis, allowing intervention when the pathology is still in an early stage and immediately directing towards a correct therapeutic approach.

Targeted metabolomics reveals the association between central carbon metabolism and pulmonary nodules

With the widespread application of low-dose computed tomography (LDCT) technology, pulmonary nodules have aroused more attention. Significant alteration in plasma metabolite levels, mainly amino acid and lipid, have been observed in patients of PNs. However, evidence on the association between central carbon metabolism and PNs are largely unknown. The aim of this study was to investigate the underlying association of PNs and plasma central carbon metabolites. We measured the levels of 16 plasma central carbon metabolites in 1954 participants who gained LDCT screening in MALSC cohort. The inverse probability weighting (IPW) technique was used to control for bias due to self-selection for LDCT in the assessed high-risk population. The least absolute shrinkage and selection operator (LASSO) penalized regression was used to deal with the problem of multicollinearity among metabolites and the combined association of central carbon metabolites with PNs was estimated by using quantile g-computation (QgC) models. A quartile increase in 3-hydroxybutyric acid, gluconic acid, succinic acid and hippuric acid was positively associated with the PNs risk, whereas a quartile increase in 2-oxadipic acid and fumaric acid was negatively associated with the risk of PNs in multiple-metabolite models. A positive but insignificant joint associations of the mixture of 16 metabolites with PNs was observed by using QgC models analyses. Further studies are warranted to clarify the association between circulating metabolites and PNs and the biological mechanisms.

Serum CHI3L1 as a biomarker of interstitial lung disease in rheumatoid arthritis

Background

Interstitial lung disease (ILD) is a relatively prevalent extra-articular manifestation of rheumatoid arthritis (RA) and contributes to significant morbidity and mortality. This study aimed to analyze the association between chitinase-3 like-protein-1(CHI3L1) and the presence of RA-ILD.

Methods

A total of 239 RA patients fulfilling the American Rheumatism Association (ACR) 1987 revised criteria were enrolled and subclassified as RA-ILD and RA-nILD based on the results of high-resolution computed tomography scans (HRCT) of the chest. The disease activity of RA was assessed by Disease Activity Score for 28 joints (DAS28) and categorized as high, moderate, low, and remission. Chemiluminescence immunoassays were applied to determine the serum levels of CHI3L1. Univariate analysis was performed and the receiver operating characteristics (ROC) curves were plotted to evaluate the correlation between RA-ILD and CHI3L1.

Results

Among the eligible RA patients studied, 60 (25.1%) patients were diagnosed with RA-ILD. Compared with RA-nILD, RA patients with ILD had significantly higher median age (median [IQR], 68.00 [62.00-71.75] vs 53.00 [40.00-63.00], p<0.001) and a higher proportion of males (21 (35.0%) vs 30 (16.8%), p=0.003). Notably, differences in DAS28 scores between the two groups were not observed. The serum level of CHI3L1 was significantly higher in RA-ILD patients (median [IQR], 69.69 [44.51-128.66] ng/ml vs 32.19 [21.63-56.99] ng/ml, p<0.001). Furthermore, the areas under the curve (AUC) of CHI3L1 attained 0.74 (95% confidence interval [CI], 0.68-0.81, p<0.001) in terms of identifying patients with RA-ILD from those without ILD. Similar trends were seen across the spectrum of disease activity based on DAS28-ESR.

Conclusion

Our findings of elevated serum CHI3L1 levels in RA-ILD patients suggest its possible role as a biomarker to detect RA-ILD noninvasively.

Acute Exacerbations of Interstitial Lung Diseases: Focus on Biomarkers

Interstitial lung diseases (ILDs) are a large group of pulmonary disorders characterized histologically by the cardinal involvement of the pulmonary interstitium. The prototype of ILDs is idiopathic pulmonary fibrosis (IPF), an incurable disease characterized by progressive distortion and loss of normal lung architecture through unchecked collagen deposition. Acute exacerbations are dramatic events during the clinical course of ILDs, associated with high morbidity and mortality. Infections, microaspiration, and advanced lung disease might be involved in the pathogenesis of acute exacerbations. Despite clinical scores, the prediction of the onset and outcome of acute exacerbations is still inaccurate. Biomarkers are necessary to characterize acute exacerbations better. We review the evidence for alveolar epithelial cell, fibropoliferation, and immunity molecules as potential biomarkers for acute exacerbations of interstitial lung disease.

A Review of Machine Learning Methods Recently Applied to FTIR Spectroscopy Data for the Analysis of Human Blood Cells

Machine learning (ML) is a broad term encompassing several methods that allow us to learn from data. These methods may permit large real-world databases to be more rapidly translated to applications to inform patient-provider decision-making. This paper presents a review of articles that discuss the use of Fourier transform infrared (FTIR) spectroscopy and ML for human blood analysis between the years 2019-2023. The literature review was conducted to identify published research of employed ML linked with FTIR for distinction between pathological and healthy human blood cells. The articles' search strategy was implemented and studies meeting the eligibility criteria were evaluated. Relevant data related to the study design, statistical methods, and strengths and limitations were identified. A total of 39 publications in the last 5 years (2019-2023) were identified and evaluated for this review. Diverse methods, statistical packages, and approaches were used across the identified studies. The most common methods included support vector machine (SVM) and principal component analysis (PCA) approaches. Most studies applied internal validation and employed more than one algorithm, while only four studies applied one ML algorithm to the data. A wide variety of approaches, algorithms, statistical software, and validation strategies were employed in the application of ML methods. There is a need to ensure that multiple ML approaches are used, the model selection strategy is clearly defined, and both internal and external validation are necessary to be sure that the discrimination of human blood cells is being made with the highest efficient evidence.

Clinically Relevant Biomarkers in Connective Tissue Disease-Associated Interstitial Lung Disease

Interstitial lung disease (ILD) complicates connective tissue disease (CTD) with variable incidence and is a leading cause of death in these patients. To improve CTD-ILD outcomes, early recognition and management of ILD is critical. Blood-based and radiologic biomarkers that assist in the diagnosis CTD-ILD have long been studied. Recent studies, including -omic investigations, have also begun to identify biomarkers that may help prognosticate such patients. This review provides an overview of clinically relevant biomarkers in patients with CTD-ILD, highlighting recent advances to assist in the diagnosis and prognostication of CTD-ILD.

Autoantibodies against citrullinated and native proteins and prediction of rheumatoid arthritis-associated interstitial lung disease: A nested case-control study

Background

To identify fine specificity anti-citrullinated protein antibodies (ACPA) associated with incident rheumatoid arthritis-associated interstitial lung disease (RA-ILD).

Methods

This nested case-control study within the Brigham RA Sequential Study matched incident RA-ILD cases to RA-noILD controls on time of blood collection, age, sex, RA duration, and rheumatoid factor status. A multiplex assay measured ACPA and anti-native protein antibodies from stored serum prior to RA-ILD onset. Logistic regression models calculated odds ratios (OR) with 95% confidence intervals (CI) for RA-ILD, adjusting for prospectively-collected covariates. We estimated optimism-corrected area under the curves (AUC) using internal validation. Model coefficients generated a risk score for RA-ILD.

Findings

We analyzed 84 incident RA-ILD cases (mean age 67 years, 77% female, 90% White) and 233 RA-noILD controls (mean age 66 years, 80% female, 94% White). We identified six fine specificity antibodies that were associated with RA-ILD. The antibody isotypes and targeted proteins were: IgA2 to citrullinated histone 4 (OR 0.08 per log-transformed unit, 95% CI 0.03-0.22), IgA2 to citrullinated histone 2A (OR 4.03, 95% CI 2.03-8.00), IgG to cyclic citrullinated filaggrin (OR 3.47, 95% CI 1.71-7.01), IgA2 to native cyclic histone 2A (OR 5.52, 95% CI 2.38-12.78), IgA2 to native histone 2A (OR 4.60, 95% CI 2.18-9.74), and IgG to native cyclic filaggrin (OR 2.53, 95% CI 1.47-4.34). These six antibodies predicted RA-ILD risk better than all clinical factors combined (optimism-corrected AUC=0·84 versus 0·73). We developed a risk score for RA-ILD combining these antibodies with the clinical factors (smoking, disease activity, glucocorticoid use, obesity). At 50% predicted RA-ILD probability, the risk scores both without (score=2·6) and with (score=5·9) biomarkers achieved specificity ≥93% for RA-ILD.

Interpretation

Specific ACPA and anti-native protein antibodies improve RA-ILD prediction. These findings implicate synovial protein antibodies in the pathogenesis of RA-ILD and suggest clinical utility in predicting RA-ILD once validated in external studies.

Funding

National Institutes of Health.

Identification of biomarkers in nonalcoholic fatty liver disease: A machine learning method and experimental study

Nonalcoholic fatty liver disease (NAFLD) has become the most common chronic liver disease. However, the early diagnosis of NAFLD is challenging. Thus, the purpose of this study was to identify diagnostic biomarkers of NAFLD using machine learning algorithms. Differentially expressed genes between NAFLD and normal samples were identified separately from the GEO database. The key DEGs were selected through a protein‒protein interaction network, and their biological functions were analysed. Next, three machine learning algorithms were selected to construct models of NAFLD separately, and the model with the smallest sample residual was determined to be the best model. Then, logistic regression analysis was used to judge the accuracy of the five genes in predicting the risk of NAFLD. A single-sample gene set enrichment analysis algorithm was used to evaluate the immune cell infiltration of NAFLD, and the correlation between diagnostic biomarkers and immune cell infiltration was analysed. Finally, 10 pairs of peripheral blood samples from NAFLD patients and normal controls were collected for RNA isolation and quantitative real-time polymerase chain reaction for validation. Taken together, CEBPD, H4C11, CEBPB, GATA3, and KLF4 were identified as diagnostic biomarkers of NAFLD by machine learning algorithms and were related to immune cell infiltration in NAFLD. These key genes provide novel insights into the mechanisms and treatment of patients with NAFLD.

Detection of WBC, RBC, and Platelets in Blood Samples Using Deep Learning

A blood count is one of the most important diagnostic tools in medicine and one of the most common procedures. It can reveal important changes in the body and is commonly used as the first stage in the process of evaluating patients' health. Even though this is a common practice, delivering examinations in laboratories can be difficult due to the availability of expensive technology that requires frequent maintenance. This study is developing an alternative deep learning computational model capable of automatically detecting cells in images of blood samples. Using object detection libraries, it was possible to train a model that was focused on this task and capable of detecting cells in images with adequate accuracy. When the identification of cells in images of blood samples was taken into account in the best results obtained, it was possible to count white cells with an accuracy of one hundred percent, red cells with an accuracy of 89%, and platelets with an accuracy of 96%, which generated subsidies to develop the primary components of a blood count. The components that were supposed to classify the various types of white cells were not carried out due to the limits of the dataset provided. On the other hand, the study can be broadened to include further works that deal with this issue because it produced satisfactory results.