Serum mass spectrometry unveil the heterogeneity of rheumatoid arthritis and reveals insights into Complement Pathways and IGHV as potential prognostic markers

Serum mass spectrometry unveil the heterogeneity of rheumatoid arthritis and reveals insights into Complement Pathways and IGHV as potential prognostic markers | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Serum mass spectrometry unveil the heterogeneity of rheumatoid arthritis and reveals insights into Complement Pathways and IGHV as potential prognostic markers Benoît Marchand, Nathalie Carrier, Elizabeth Beaulieu, Dominique Lévesque, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9359440/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 6 You are reading this latest preprint version Abstract Objectives We aim to identify novel biomarkers of rheumatoid arthritis (RA) outcomes by performing serum proteomic profiling of our longitudinal RA cohort using Data-Independant Acquisition (DIA) mass spectrometry. Methods. Serum proteomes from 107 previously untreated early RA patients were recruited in the EUPA cohort and their sera were analyzed at baseline and at the 12-month follow-up visit using DIA mass spectrometry technology. Clustering analyses on both baseline and follow-up serum profiles was performed to assess overall patient heterogeneity. Generalized estimating equations (GEE) analyses of longitudinal serum proteome data was used to query for proteins associated with disease activity or erosion outcomes. Functional networks of the identified proteins were explored using KEGG enrichment analysis while novel predictors of RA outcomes were identified using generalized linear models (GLM) with baseline serum proteomes, and performance was evaluated by receiver operating characteristic (ROC) curves. Results. RA patients could be separated in 2 distinct clusters based on their serum proteome profiles, regardless of disease activity or erosion outcomes. Compared to CRP, the protein signature composed of APOC4 and SAA1/SAA2 exhibited improved predictive performance for disease activity. Furthermore, an additional protein signature combining CTBS and IGHV5-10-1/IGHV5-51 outperformed classic autoantibodies serology status to predict erosiveness. Functional network enrichment analyses uncovered association between the complement system and RA progression. Conclusion. Leveraging serum proteomic profiling, we identified novel biomarkers of RA outcomes, reinforcing the notion that protein signatures can improve predictive performance while highlighting crucial elements of pathophysiology that might lie in the complement system. Further clinical validation of these predictors may potentially pave the way toward improved personalized treatment strategies and ultimately better RA management. Mass Spectrometry Rheumatoid Arthritis Proteomics RA biomarker Erosion Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction: Rheumatoid arthritis (RA) is the most common type of autoimmune inflammatory arthritis, affecting approximately 1% of the global population. Current classification of RA still relies heavily on clinical manifestations combined with serological analysis of autoantibodies, primarily rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPA), and erosiveness. Although the heterogeneity of RA pathogenesis between RF/ACPAs positive and negative serotypes has been highlighted at the level of genetic predisposition 1 , immune cell profile 2 and disease aggressiveness including risk of developing bone erosion 3 , 4 , these differences are too subtle to readily inform clinical decision. Furthermore, they are inadequately addressed by the presently available clinical tools. Current management strategies for RA thus primarily aim to provide early and aggressive treatment 5 , 6 in order to quickly control inflammation and prevent irreversible joint damage that could lead to disability. Many disease-modifying antirheumatic drugs (DMARD) are approved for treating RA, with methotrexate being the gold standard and often considered as the first-line therapy for DMARD-naive patients 7 . Indeed, RA management guidelines by the American College of Rheumatology (ACR) recommend a treat-to-target strategy 6 , which relies on clinicians’ ability to regularly assess disease activity, monitor treatment response and adjust treatment until remission is reached. In order to make such decisions and to monitor disease activity, several tools are available such as the CDAI, SDAI and DAS28-CRP which are all recommended by the ACR 8 . However, disease activity still does not accurately predict erosion risk 9 , especially in patients with low disease activity, where approximately 15% still develop erosion 10 , 11 , likely due to underlying subclinical inflammation 12 . Furthermore, these disease activity measures fail to reliably predict response to treatment 13 , 14 . To circumvent these imperfections in the assessment of RA patients, a matrix estimating erosion progression risk (AUC 0.68) was developed using four baseline characteristics (RF positivity, X-ray erosions, serum C-reactive protein [CRP] levels, and Swollen Joint Count 28); however this tool still lacks sufficient predictive power (PPV 0.30) for integration into clinical practice 15 . Therefore, identifying biomarkers able to predict treatment response, erosiveness, or joint damage would allow for more accurate risk depiction and better support clinical decisions in managing RA. The multi-biomarker disease activity (MBDA) score is a 12-protein biomarker test that correlates with clinical disease assessment using the DAS28-CRP 16 , 17 . While CDAI, SDAI and DAS28-CRP scores rely on assessment of RA signs and symptoms, the MBDA offers a more objective evaluation of disease activity, free from the biases of physical examinations and inter-user variability. MBDA scoring has been suggested to correlate with radiographic progression with improved performance when adjusted for age, sex and leptin (a peptide hormone regulating energy balance and body weight) levels 9 , 18 – 20 . Nonetheless, baseline MBDA scores have not been shown to reliably predict treatment responses, remission or radiographic progression 14 , 21 , leaving its cost-effectiveness and clinical utility unsettled. Interestingly, a recent ambitious plasma proteomic study, in 41,931 participants from the UK Biobank Pharma Proteomics Project, demonstrated that combining proteomic signatures, comprising 5 to 20 proteins, with clinical features could significantly improve risk prediction for various diseases 22 . Specifically, protein signatures improved risk prediction in rheumatic diseases such as Gout (ΔC-index: 0.0307, Integrated Discrimination Improvement (IDI): 1.3%), Sjögren’s (ΔC-index: 0.1178, IDI: 11.78%) and Osteoarthritis (ΔC-index: 0.0235, IDI: 2.5%). This finding supports the potential of proteomic investigations endeavours towards the discovery of protein signatures predictive of RA outcomes. Several proteomic studies have previously investigated biomarkers associated with RA diagnosis 23 – 25 or in the pre-clinical RA phase, i.e. in at-risk and pre-progression individuals 26 , 27 . Previous work from our group has shown that adding high serum levels of 14-3-3η to autoantibodies (ACPAs) and CRP slightly improved prediction of impending rapid erosion progression in our longitudinal Early Undifferentiated PolyArthritis (EUPA) cohort 28 . Moreover, our group provided proof-of-concept evidence that miRNA profiles derived from PBMC-generated osteoclasts can enhance the identification of rheumatoid arthritis (RA) patients at risk of developing erosions, with predictive models achieving an accuracy of 78% and an area under the curve (AUC) of 0.85; however, this approach is not yet clinically available 29 . In the current study, we leveraged the superior dynamic range and deep coverage of Data-Independent Acquisition (DIA) mass spectrometry 30 to overcome the analytical challenges associated with blood proteomic analysis 31 , enabling us to comprehensively profile the serum protein landscape in untreated and treated RA patients. We took an exploratory approach to identify serum protein biomarkers with potential to predict RA disease activity and erosion outcomes. Material and methods: Patient Cohort: Patients were selected from the longitudinal Early Undifferentiated PolyArthritis (EUPA; NCT00512239) cohort at the CIUSSS de l’Estrie-CHUS, Canada. This study used serum samples collected at the baseline and the 12-month follow-up visits from previously DMARD-naïve RA patients enrolled in the EUPA cohort described elsewhere 32 , 33 . Patients were selected based on their disease activity status at 12 months, with DAS28-CRP scores ≥ 3.2 indicating active disease and ≤ 2.6 indicating remission 8 . Additionally, we considered the erosion outcome at the 30- to 60-month visit, categorizing patients as erosive (Sharp erosion score ≥ 5) or non-erosive (Sharp erosion score < 5) 34 . Patients were extracted from our biobank in order to equilibrate both seropositivity and persistence of activity at the 12-month follow up visit between the two groups. The initial patient subset consisted of CCP+/RF+ individuals, divided into four groups of 10 (n = 40): active/erosive, active/non-erosive, remission/erosive, and remission/non-erosive (Supp. Figure 1 ). A subsequent batch included 22 patients who were either CCP + or RF+, with 12 in an active state and 10 in remission; the remaining 50 participants were seronegative (CCP-/RF-), comprising 25 active and 25 in remission. Patient selection was extended to seronegative patients to broaden the scope of our analysis to identify biomarkers of severity independent of serology status. Indeed, our group has previously shown that serological status has lost most of its predictive value in modern cohorts of RA 33 , hence the need for more reliable biomarkers and the expended analysis to seronegative patient. In addition, this extension allowed us to explore potential differences in serum proteome specific to seropositive or seronegative patients. Serum samples were processed following standard operating procedures of the EUPA cohort. Briefly, blood collected from individuals was kept at 4°C (up to a maximum of 5 days) until samples were processed; serum was then prepared, aliquoted, and frozen at -20°C 35 . Mass Spectrometry (MS): Sample preparation: Prior to serum processing for mass spectrometry analysis, the protein content was quantified using BioRad Protein Assay Dye (#5000006) and 50 µg of patients' serum protein were resuspended in Urea buffer (8 M urea, 10 mM Hepes, 1 M NH4HCO3) to a final volume of 50 µl. Samples were processed as follows: proteins were first reduced by adding DTT to a final concentration of 5 mM, followed by heating for 2 min at 95°C and an additional incubation at room temperature (RT) for 30 min. The reduced protein samples were then alkylated by adding chloroacetamide to a final concentration of 7.5 mM and incubating for 20 min in the dark at RT. Subsequently, proteins were digested overnight with 1 µg of trypsin (Promega; V5280) at 30°C. The following day, trypsin digestion was stopped by addition of trifluoroacetic acid (TFA) to a final concentration of 0.2% and the generated peptides were desalted and cleaned using ZipTip C18 columns (Millipore Sigma). Peptide desalting was performed as follows: ZipTips pre-wetted in 100% acetronitrile (ACN) were equilibrated with 0.1% TFA, then samples were loaded by repeated up and down pipetting, washed with 0.1% TFA, and eluted in 1% formic acid (FA)/50% ACN. Desalted samples were then vacuum dried and resuspended in 1% FA. Data acquisition: For each sample, 250 ng of peptides were injected in a HPLC (nanoElute, Bruker) and loaded onto a trap column at a flow of 4 µl/min (Acclaim PepMap100 C18 column, 0.3 mm id × 5 mm, Dionex Corp.) then eluted onto an analytical C18 column (1.9 µm beads, 75 µm × 25 cm, PepSep). Peptides were eluted over a 2-h gradient of acetonitrile (5–37%) in 0.1% FA at 400 nl/min, while being injected into a TimsTOF Pro ion mobility mass spectrometer equipped with a Captive Spray nano electrospray source (Bruker). Data were acquired as follows: for each single TIMS (100 ms) in data-independent acquisition and parallel accumulation serial fragmentation (diaPASEF) mode, we used one mobility window consisting of 27 mass steps (m/z between 114 to 1414 with a mass width of 50 Da) per cycle (1.27 sec. duty cycle). These steps cover the diagonal scan line for + 2 and + 3 charged peptides in the m/z-ion mobility plane. Protein identification and quantification: Protein identifications were performed on raw TimTOF data using DIA-NN v1.8.1 software 36 and the Uniprot human reference proteome (UP000005640_9606). An in silico spectral library was generated from peptides from our DIA runs. Default DIA-NN settings were used with the following parameters: 1 missed cleavage allowed, cysteine carbamidomethylation was set as fixed modification, Mass and MS1 accuracy were set at 20 ppm, and Min/Max precursor charge were set at 2 and 4. The MaxLFQ-based quantification was performed with a precursor q value filtering at 0.05 37 . Data processing: We used ProStaR/DAPAR (Wieczorek S. et al. 2017) software 38 and loaded the Label-Free Quantitation (LFQ) protein intensities obtained from DIA-NN. The LFQ intensities were log2-transformed and filtered to retain only proteins identified by 2 or more unique peptides. Only the proteins that had less than 30% of Partially Observed Values (POV) in both baseline and 12 months post-treatment condition were kept. Then, normalization was performed by quantile centering (15% quantile) and missing values were imputed using DetQuantile algorithm for both POV and Missing in the Entire Condition (MEC) with Quantile = 5% and Factor = 0.2. Statistical analysis: R software v4.3.2 39 was used to perform data processing, statistical analyses and visualization, with the following packages: cluster, cowplot, dplyr, ggforce, ggfortify, ggplot2, ggrepel, ggsankey, gt, purrr, stringr, tibble, and tidyr. Patient clustering analysis. We performed two separate clustering analyses to assess unbiased RA patients’ aggregation based on their serum proteome profiles at baseline and 12-month follow-up visit. The optimal number of clusters was evaluated using Silhouette width and a consensus of multiple methods. We proceeded using the Hierarchical Clustering on Principal Components (HCPC) method of the FactoMineR and factoextra R packages 40 , 41 . Briefly, Principal Component Analysis (PCA) analyses were conducted on both baseline and post-treatment serum proteome data, followed by HCPC clustering using Euclidean distance metrics and Ward’s method, with a maximum of 100 iterations for consolidation. Differently abundant proteins (DAP) between patient clusters were identified by Mann-Whitney U test with Bonferroni correction. The clinical and demographic patient data comparisons between clusters were conducted using the Mann-Whitney U test for continuous values or either the Chi-square test of independence or Fisher’s exact test for categorical values. Protein levels were normalized using z-scores calculated for each protein as (value – mean of all patients)/standard deviation of all patients, for baseline and 12-month proteomes. This allowed comparison of multiple protein serum levels across the patient cohort. Longitudinal serum proteome analysis. For repeated measure analysis, Generalized Estimating Equations (GEE) models were performed using the geepack R package 42 . GEE models were used to evaluate association of serum proteins levels (baseline and post-treatment) with binomial outcomes of disease activity (DA) or erosion status at the 12- to 30-month visits as the response variables. A GEE model was performed for each protein as a predictor term with the following additional predictors as potential confounders: age, gender, serology and symptom duration. The blood draw timeline (baseline vs post-treatment) was used as the repeated measure (wave) with a binomial family of distribution. GEE models performed for DA outcomes used an “exchangeable” (compound symmetry) correlation structure to leverage DA outcomes measured both at baseline and post-treatment at the 12 months visit, while an “independent” structure was used for analysis of erosion outcome. GEE analyses were repeated using serology (ACPA/RF)-specific patient subgroups or all patients combined. However, a limited number of patients were excluded from GEE models due to the following missing values: symptom duration NA (1), Sharp score NA (1), baseline DAS28-CRP NA (4), and baseline DAS28-CRP between 2.6–3.2 (3). The Odds Ratios (OR) with a 95% Confidence Interval (CI) and p values were extracted from the GEE models using easystats package 43 and visualized in volcano plots. Baseline serum proteome analysis. Logistic regression models were conducted to identify potential biomarkers predictive of DA or erosion outcomes using the generalized linear models (glm) function with a binomial ("logit") family of distribution from the “stats” R package. Models were performed with baseline serum protein levels as predictor terms, also including age, gender, serology, and symptom duration confounders as additional terms for DA or erosion binomial outcomes as response terms. The Odds Ratios (OR) with a 95% Confidence Interval (CI) and p values were extracted from the models using easystats package, then visualized in Forest plots. Subsequently, the selection of predictive features was refined by filtering proteins which had fold changes (FC) above 1.25X or below 0.75X in the baseline serum proteome. Model performance assessment Receiver Operating Characteristic (ROC) curve analyses were performed on logistic regression-selected proteins to evaluate their ability to discriminate between DA or erosion outcomes. ROC curves were analyzed using the predicted probabilities from univariate and multivariate logistic regression models using the pROC R package 44 . For multivariate logistic regression models, combinations of predictor terms up to 4 for DA or 3 for erosion (due to lower n) from selected proteins were tested, with additional terms for age, serology, and symptom duration confounders. Predictor combinations were filtered using ROC curve AUC values to retain only those combinations where the addition of a predictor term increased AUC values by a minimum threshold of 0.01. To further refine our results, models were filtered using a penalized-AUC value based on the number of predictors to select the most optimal predictor combinations. The penalized-AUC value was calculated using a threshold of 0.04 per additional predictor terms. Models were ranked based on their penalized AUC and positive predictive values (PPV) values, which were weighted equally (1:1), and the top-performing models were selected. DeLong’s test was used to compare the performance of the identified predictors, allowing to determine if they contributed statistically significant improvements. KEGG pathway enrichment analysis. KEGG pathway enrichment analyses were carried out with active subnetworks using the PathfindR R package 45 , 46 , which leverage Protein-protein Interaction Network (PIN) information. KEGG enrichment analyses were performed using proteins and p values from our GEE results as inputs, the KEGG gene set, the Biogrid PIN, the Benjamini & Hochberg p adjustment and enrichment threshold of 0.05 which was repeated over 10 iterations. Similarly, KEGG pathway analysis was also conducted on DAPs between patient clusters using Bonferroni adjusted p values from Mann-Whitney U test and log2 fold changes. Results: Cohort description and experimental design Serum proteomic profiling was performed on a selection of 48 seronegative and 59 seropositive RA patients previously enrolled in the EUPA longitudinal cohort 33 between 2005 and 2019, described in Table 1. The selected patients were all DMARD-naïve at inclusion and consisted of individuals with disease activity (DA) levels at the 12-month visit assessed as active or in remission in a balanced ratio. The initial selection of 40 seropositive patients incorporated an additional criterion of erosive status at the 30-60-month visits, however, this criterion was removed for subsequent patient selections to expand the cohort (Supp. Figure 1 ). The demographics, including age, sex, body mass index (BMI), were similar among our seronegative and seropositive patient subgroups. RA patients from both groups showed comparable baseline disease activity (SJC68, TJC66, DAS28-CRP, SDAI, CDAI). By the 12-month time point, seropositive and seronegative patients had been exposed to DMARDs for a median of 6.8 months (IQR; 4.9–8.5) and 7 months (IQR; 5.2–8.8), respectively. DIA mass spectrometry serum profiling enabled identification of 869 proteins from which 368 were retained following quality control and filtering (see Methods) for downstream statistical analyses. Exploring RA patients clustering based on serum proteome landscapes. We began by evaluating whether serum proteome profiles could identify distinct RA patient subgroups, using clustering analyses at baseline and post-treatment (12-month visit). There were no confounding effects of therapy at baseline, while the 12-month visit accounted for longitudinal changes of protein expression patterns under treatment. Hopkins’ statistic values of 0.666 when using baseline serum proteome and of 0.665 for the post-treatment serum indicated only moderate clustering tendencies. We thus applied Hierarchical Clustering on Principal Components (HCPC) to separate patients in two clusters based on serum proteomic profiles at baseline (Fig. 1 A) and post-treatment (Fig. 1 B). This clustering reflected the optimal number of clusters suggested by the silhouette width (Supp. Figure 2 A,C), as well as by the majority of clustering metrics (Supp. Figure 2 B,D). Comparison of clinical and demographic features across clusters from both baseline and post-treatment proteome clustering did not reveal significant associations with clinical manifestations or outcomes (Table 2 ). Nonetheless, cluster B1 exhibited a higher proportion of patients seropositive for either RF or ACPAs (69%) compared to cluster B2 (43%, p = 0.011) (Table 2 ; Supp. Figure 3 A). Disease activity measures at the 12-month visit, such as the DAS28-CRP, SDAI and HAQ, were comparable across clusters at baseline (Supp. Figure 3 C) and post-treatment (Supp. Figure 3 D). Furthermore, patients across clusters had similar CRP blood levels, a surrogate of inflammation, both at baseline or at the 12-month visit (Table 2 ). There was also no clear association of DA (Supp. Figure 3 E-F) or erosion (Supp. Figure 3 G-H) outcomes with patients clustering. To further investigate our clustering results, we analyzed the differently abundant proteins across patient clusters by performing KEGG enrichment analysis. Notably, the complement and coagulation cascades were the most significantly enriched pathways both in baseline (Supp. Figure 4 A) and post-treatment (Supp. Figure 4 B) clustering. Patients’ serum levels of proteins related to the complement and coagulation cascades pathway, as shown by the mean z-scores, segregated both at baseline (Fig. 1 C; Supp. Figure 5 ) and post-treatment (Fig. 1 D; Supp. Figure 6). Next, we compared patients’ trajectories between baseline serum-based clusters and corresponding post-treatment serum clusters (Fig. 1 E). Patients in cluster B1 were split 43% in P1 and 57% in P2, while most patients (88%) of cluster B2 were assigned to P2. Notably, we found significant differences in mean z-scores for complement-related proteins between baseline and 12-month serum samples of patients who transitioned from B1 to P2 (p adj. = 2.8e-5) and B2 to P1 (p adj. = 0.016) clusters (Fig. 1 F). Identification of disease activity associated proteins by longitudinal analysis of RA patient serum proteomes. For a longitudinal assessment of proteins correlation with outcomes, we performed Generalized Estimating Equation (GEE) models incorporating both baseline and 12-month measurements. By analyzing both time points simultaneously, we captured the dynamic changes in protein expression that may be associated with disease activity or erosion outcomes. Studies on prognosis of RA patients from the Canadian Early Arthritis Cohort (CATCH) suggested seropositive RA patients showed lower treatment responses and higher risk of erosion 3 . Thus, we first implemented GEE models separately on seropositive and seronegative patient subsets for each protein with binomial outcomes, then GEE analyses were repeated on complete cohort data. First, GEE analyses using serum proteome with DAS28-CRP based binomial outcome identified 39 and 75 disease activity-associated proteins (p < 0.05) in seronegative (Fig. 2 A) and seropositive (Fig. 2 B) subgroups, respectively. Furthermore, analysing the complete RA cohort’s serum proteomes regardless of serology revealed 64 proteins significantly correlated with DA (p < 0.05) (Fig. 2 C). A subset of these proteins exhibited FCs above 1.25X or below 0.75X at either baseline or 12-month, thus most likely to have significant impact, including ADIPOQ, CRP, FAH, IGHV3-43, IGKV2-29; IGKV2D-29, IGLV1-36, IGLV3-27, IGLV8-61, SAA1, SAA2, TAGLN2 and XIRP2. Moreover, there was partial overlap between proteins identified from analyses performed using serology-specific or the full cohort data, with 14 proteins consistently identified across all three GEE analyses (Fig. 2 D). Subsequently, we assessed biological function by conducting KEGG pathway enrichment analysis using DA-associated proteins identified through our GEE analysis of serum proteomes from the full cohort. The top 10 enriched KEGG terms related to RA disease activity were primarily components of the complement system, with the addition of lipoproteins and platelet activation (Fig. 2 E). Identification of erosion associated proteins by longitudinal analysis of RA patient serum proteomes. We applied the same GEE approach to analyze serum proteins associated with erosion status binomial outcomes 47 . These GEE analyses of RA patients’ serum proteomes identified 22 and 44 proteins associated with erosion (p < 0.05) in seronegative (Fig. 3 A) and seropositive (Fig. 3 B) subgroups, respectively. Additionally, 27 proteins were identified as significantly correlated with erosion (p < 0.05) by analysing the complete RA cohort’s serum proteomic data regardless of serology (Fig. 3 C). Among these, CA1, CETP, CTBS, DNAH11, ENPP2, IGHV1-18, IGHV1-2, IGHV1-46, IGHV2-26, IGHV5-10-1, IGHV5-51, IGLV3-9, MPO exhibited FCs ≥ 1.25X or ≤ 0.75X at either baseline or 12-month. Overall, there was more limited overlap between erosion-associated proteins identified across GEE analyses using serology-specific subgroups or the full cohort (Fig. 3 D), and none was common in the 3 GEE analyses. Nonetheless, the analysis of KEGG pathways using erosion-associated proteins identified from the complete cohort suggested enrichment of phagosome and neutrophil extracellular trap formation pathways (Fig. 3 E). Exploring baseline serum profiles to predict RA outcomes. To explore biomarkers that may predict rheumatoid arthritis outcomes, including DA at 12-month and erosiveness, we analyzed baseline serum proteomic profiles using generalized linear modeling (GLM) with binomial outcome (logistic regression) approaches, adjusted for age, gender and serology as confounders. For DA outcome at 12-month, we identified eleven proteins with baseline serum levels significantly associated (p < 0.05) (Fig. 4 A). Within these proteins, APOC4, PFN1, and SAA1;SAA2 had FCs ≥ 1.25X or ≤ 0.75X (in blue in Fig. 4 A). These findings indicate that baseline serum concentrations of these proteins may serve as predictors of resistance to first-line therapies. Next, we analyzed the association between baseline serum protein levels and binomial erosion outcome using GLM models adjusted for age, gender and serology. We observed 17 proteins with baseline serum levels significantly (p < 0.05) associated with erosion (Fig. 4 B). The following 7 proteins had FCs ≥ 1.25X or ≤ 0.75X : CA1, CTBS, IGHV1-18, IGHV1-46, IGHV2-26, IGHV5-10-1;IGHV5-51, and MPO (in blue Fig. 4 B). Evaluating our protein biomarker candidates’ performance To further evaluate the performance of the identified protein biomarker candidates, we conducted ROC curve analyses to assess their discriminatory power in distinguishing between patients based on 12-month DA or erosion outcomes. For DA, each protein identified (Fig. 4 A) performed non-inferiorly compared to CRP, a known predictor (Fig. 5 A, see also Supp. Tables 1 and 2 ); SAA1;SAA2 (AUC = 0.636, PPV = 59.6%) and PFN1 (AUC = 0.623, PPV = 60%) had slightly higher AUC values compared to CRP (AUC = 0.601, PPV = 59.1%), a non statistically significant difference. Next, we explored whether combining multiple protein predictors (up to four only to avoid overfitting) could enhance overall performance. The ROC curves from our top five models selected based on balanced improvements of AUC and PPV are presented in Fig. 5 B and their performance were compared using DeLong’s test (Supp. Tables 1 and 2 ). The predictive performance was further improved when combining proteins, such as SAA1;SAA2 with PFN1 (AUC = 0.677, PPV = 71.1%) or adding APOC4 to this combination (AUC = 0.691, PPV = 65.9%). Neither the SAA1;SAA2-PFN1 (p = 0.13) nor the SAA1;SAA2-PFN1-APOC4 (p = 0.09) signatures showed significant superiority over CRP alone (Fig. 5 B, Supp. Tables 1 and 2 ). Next, we assessed the performance of our protein candidates identified in Fig. 4 D as predictors of erosiveness. Notably, autoantibody serology, a known predictor of erosion 4 , 18 , 48 , exhibited a ROC AUC value of 0.722 (PPV = 50%) when adjusted for age and symptom duration as confounders (Fig. 5 C, Supp. Table 1). Interestingly, all protein predictors of erosion tested yielded AUC values higher than serology (Supp. Table 1), but only IGHV1-18 (AUC = 0.832, PPV = 61.9%, p = 0.009) was statistically superior (Supp. Table 2 ). The ROC curves from the adjusted univariate GLM models from the top 5 proteins with FCs ≥ 1.25X or ≤ 0.75X, along with serology (RF or ACPAs), are shown in Fig. 5 C. Three of these five proteins are related to variable domains of immunoglobulin heavy chains. Given our limited number of erosive patients, we combined only up to 3 protein predictors. The results from predictor signatures were refined by selecting the top 5 protein combinations with the best performance based on both their AUC and PPV values (Fig. 5 D). Combining CTBS with IGHV1-18 (AUC = 0.846, PPV = 68.2%) significantly outperformed serology (p = 0.004). Although the CTBS-IGHV1-18 signature had slightly higher PPV compared to IGHV1-18 alone (68.2% vs 61.9%), their AUCs were not statistically different (0.846 vs 0.832, p = 0.255). While improved performance over serology was observed for several protein signatures (Supp. Table 2 ), these failed short of exhibiting improvements over IGHV1-18 alone. Discussion: Predicting which very early RA patients will respond to specific therapies or will progress to joint erosiveness remain some of the major challenges in contemporary rheumatology, yet are essential for truly personalized RA management. While we were able to capture RA patient heterogeneity through serum proteomes both at baseline and at the 12-month follow-up, clustering patients did not correlate well with clinical outcomes (DAS28-CRP or erosion). This highlights the observable heterogeneity in RA, but also underscores the current limitations in translating proteomic diversity into actionable clinical insights. RA patients in stable remission from the RETRO cohort could be segregated based on proteomic profiles into 4 clusters 49 . Similar to our observations, their patient clustering did not clearly associate with future RA outcomes but rather with clinical features. Our current methodology did not allow to assess the impact of these clusters on response to specific treatments. Across studies, the number and the degree of separation of patient clusters using plasma/serum proteomes have greatly varied. Several studies even noted an overlap between clusters of RA patients with clusters composed of healthy controls 23 , 50 , while others were able to observe sharper cluster separation 25 , such as in Tocilizumab responders and non-responders 51 . A more comprehensive analysis, including both distal (serum/plasma) and local (synovial) samples, alongside multi-omic approaches may help refine our understanding of RA patient phenotypes, improve clustering and help establish a link to relevant clinical outcomes. In this study, differently abundant serum proteins between patient clusters were enriched for proteins of the complement pathways (Supp. Figure 4 ), despite the absence of association with serology or DA outcomes (Supp. Figure 3 ). CRP blood levels also did not distribute across patient clusters defined by baseline (Supp. Figure 7A) or 12-month (Supp. Figure 7B) proteomic profiles. This suggests that heterogeneity of circulating complement components is not merely a reflection of inflammation levels but reflects a different polarisation of the immune system in some RA patients. While we did not observe a clear link between serum complement proteins and serology, recent single-cell RNA sequencing data revealed complement cascade gene enrichment in macrophages of ACPA negative compared to ACPA positive patients 52 . While further work is needed to elucidate a potential serology-specific contribution of the complement pathways, its role in the pathogenesis of RA was previously documented. Previous studies revealed complement pathway terms were enriched in RA patients compared to healthy individuals 24 , 25 , associated with DAS28 in stable RA patients 49 , and associated with Tocilizumab response 51 . Furthermore, several complement deficiency mouse models demonstrated contribution of the complement pathways to collagen (CIA) and collagen antibody (CAIA) induced arthritis 53 – 57 . In addition, rheumatoid factor is known to activate complement 58 , 59 , while C5a, an anaphylatoxin, seems to be the primary complement activation product driving tissue damage in RA, though membrane attack complex deposition 60 and C3b-mediated opsonization 61 also play significant roles. A more comprehensive characterization of complement proteins involved in RA by integrating analysis of the synovium, independently of serological profiles, may offer valuable insights for identifying patients who could benefit from complement system modulators such as avacopan, a C5aR1 antagonist 62 , highlighting complement modulation as an emerging therapeutic option under investigation for rheumatic diseases 63 . Our cohort of RA patients had similar disease activity at baseline, decreasing variability in protein expression and allowing the identification of subtler changes. Consequently, we did not apply false discovery rate (FDR) correction in GEE and GLM analyses using binomial outcomes, enabling identification of a broader set of proteins more suitable for network enrichment analysis. While this exploratory approach increases the risk of false-positives, 29 proteins remained significantly associated with DA outcome after applying Storey's q-value FDR correction (Supp. Figure 8A) 64 . These proteins were also consistently enriched for similar KEGG pathways (Supp. Figure 8B). In contrast, only IGHV1-18 was retained as a significant predictor of erosion outcome following FDR correction, likely due to reduced statistical power from smaller and imbalanced comparison groups of erosive patients. Our longitudinal analysis aligns with previous reports showing elevated serum levels of inflammation markers, CRP, SAA1 and SAA2 in RA patients and their association with disease progression 23 , 65 – 68 . Perhaps not surprisingly, CRP and SAA, both components of the MBDA score, have been associated with remission, thus underscoring their relevance in RA pathogenesis 16 , 69 . We also observed association of LRG1 serum levels with disease activity, validating recent reports characterizing LRG1 as a novel inflammation marker correlating with DAS28-CRP/ESR 70,71 . Several proteins associated with DA also correlated with CRP levels, a surrogate marker of inflammation, though with varying strengths and directions (Supp. Figure 9). Among the other proteins associated with DA outcome, several are linked to the complement system from the Lectin (FCN3, MASP1) and the classical (C1QB) pathways as well as with the common effector C9. However, circulating levels of complement components by themselves may not fully inform on their activation and pathogenic contribution, particularly within the joint tissues. Analyses focused on the baseline serum proteomic profiles of treatment-naïve RA patients revealed that APOC4, PFN1, and SAA1;SAA2 exhibited predictive capabilities for the 12-month DA outcomes. Combining these proteins showed a trend toward improved performance over CRP, supporting the potential for serum protein signatures to improve current clinical biomarkers. We also identified novel potential predictors of RA erosiveness, including baseline serum levels of MPO and several immunoglobulin variable chain proteins, particularly IGHV1-18. Immunoglobulin chains, including IGHV1-18, have been previously reported as enriched in RA patient exosomes, highlighting their putative role in RA 72 . Recently, the frequency of usage of several V/D/J genes of IGHV was reported to associate with disease activity in RA and SLE 73 . Their relative overabundance might suggest that a group of patients develops an oligoclonal response to one or a small subset of autoantigens. The target antigen(s) of these clones remains unknown but might be a subset of citrullinated proteins, such as citrullinated vimentin, whose corresponding autoantibodies are associated with aggressive course of RA 32 . Taken together, these findings suggest that the identification of patients at risk of poor treatment responses or erosive disease through protein signatures may be possible. Moreover, the identification of the targeted antigens might lead to a more precise understanding of RA pathophysiology and to the development of better diagnostic and prognostic tools. The presence of IGHV1-18, or potentially the preferential use of other specific B cells receptor chain, may identify a subset of B cell receptor clones that preferentially recognize autoantigens in patients with erosive rheumatoid arthritis. This observation could suggest a novel framework for understanding seropositivity, one that focuses on a subset of B cells predisposed for autoreactivity, independent of their classical targets such as ACPA or carbamylated proteins. Our DIA mass spectrometry study design intentionally avoided the removal of highly abundant proteins to maximize the number of patients analyzed longitudinally. Consequently, this study may be limited by its depth of serum proteomic profiling, particularly for low-abondance proteins, due to the challenges with mass spectrometry analyses of high dynamic ranges of plasma and serum protein levels 74 . Also, further validation of the identified outcome predictors in an independent cohort is required. Optimised protein signatures have the potential for early tailoring of treatment strategies for RA patients at risk of aggressive disease progression and ultimately improve their quality of life. Conclusion: This longitudinal serum proteomic profiling of 107 patients led to the identification of novel protein candidates as potential biomarkers of RA outcomes. We analyzed baseline serum proteome from treatment-naïve RA patients, thereby avoiding any confounding effects of DMARDs on serum protein levels. While these analyses enabled clustering of patients based on their serum proteomic profiles, we were unable to establish clear links between patient clusters and clinical outcomes. Nonetheless, we identified protein signatures that outperformed classic serologic prediction of RA erosiveness. These findings highlight the complexity and heterogeneity of RA and support the need for further in-depth characterization using multi-omic approaches. Abbreviations: ACN acetonitrile ACPA anti-citrullinated protein antibodies ACR American College of Rheumatology AUC area under the curve BMI body mass index CAIA collagen antibody–induced arthritis CATCH Canadian Early Arthritis Cohort CCP cyclic citrullinated peptide CDAI clinical disease activity index CHUS centre hospitalier universitaire de Sherbrooke CIA collagen-induced arthritis CIUSSS centre intégré universitaire de santé et de services sociaux CRP C-reactive protein DAP differentially abundant proteins DAPAR differential analysis of protein abundance with R DAS28-CRP disease activity score 28 using CRP DIA data-independent acquisition diaPASEF data-independent parallel accumulation–serial fragmentation DMARD disease-modifying antirheumatic drug DTT dithiothreitol ESR erythrocyte sedimentation rate EUPA Early Undifferentiated PolyArthritis cohort FDR false discovery rate GEE generalized estimating equations GLM generalized linear model HAQ health assessment questionnaire HCPC hierarchical clustering on principal components HPLC high-performance liquid chromatography IGHV immunoglobulin heavy chain variable region IQR interquartile range KEGG Kyoto Encyclopedia of Genes and Genomes LFQ label-free quantification m/z mass-to-charge ratio MBDA multi-biomarker disease activity MEC missing in entire condition miRNA microRNA PBMC peripheral blood mononuclear cells PCA principal component analysis PIN protein interaction network POV partially observed values PPV positive predictive value RA rheumatoid arthritis RF rheumatoid factor RNA ribonucleic acid ROC receiver operating characteristic SAA serum amyloid A SDAI simplified disease activity index SJC68 swollen joint count 68 SLE systemic lupus erythematosus TFA trifluoroacetic acid TIMS trapped ion mobility spectrometry TJC66 tender joint count 66 Declarations: Ethics approval and consent to participate: Patients enrolled in the Early Undifferentiated PolyArthritis (EUPA) cohort gave their informed consent and study was approved by the Ethics Review Board of the CIUSSS de l’Estrie-CHUS (ClinicalTrials.gov ID: NCT00512239). Consent for publication: Not applicable Availability of data and materials: The proteomic datasets generated during this study are not publicly available because patient consent for sharing individual proteomic data was not obtained. Competing interests: GB declares an advisory relationship with Otsuka Canada, and received honoraria for presentations and unrestricted research funding from Biocon Biologics Canada Inc. SR declares an advisory relationship with Amgen Canada, Kyowa Kirin, and Apotex. SR received speaker fees and funding from Kyowa Kirin, as well as funding from Insmed. HAC declares an advisory relationship with Abbvie, Amgen Canada, AstraZeneca, Celltrion, Eli Lilly & Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Novartis Canada, Pfizer Canada, Sandoz Canada and Sobi. HAC received speaker fees from Abbvie, Amgen Canada, AstraZeneca, Bristol Myers Squibb, Celltrion, Eli Lilly & Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Mantra Pharma, Novartis Pharmaceuticals Canada, Pfizer Canada and Sobi. HAC was awarded funding from Abbvie, AstraZeneca, Bristol Myers Squibb, Daiichi Sankyo, Eli Lilly & Co, Fresenius Kabi USA, Neomed, Novartis Pharmaceuticals Canada, Pfizer Canada, Sanofi and Viela Bio. All other authors declare no known competing interests related to the work reported in this article. Funding: The EUPA cohort was supported by the Canadian Institutes for Health Research MOP-110959. We also acknowledge previous support from The Arthritis Society Grants 00/201 and RG06/108. From July 2007 to May 2025, the EUPA cohort also received funding from the Canadian ArThritis CoHort. SR, GB and HAC are part of the Centre de Recherche du Centre Hospitalier Universitaire de Sherbrooke (CR-CHUS), which received a team funding from the Fonds de Recherche du Québec–Santé (FRQS). FMB, SR, GB and HAC received support from an internal grant from CR-CHUS for Projet structurants en recherche translationnelle. HAC is a Clinical Research Scholar – Junior 2 from FRQS (https://doi.org/10.69777/369863). HAC is the current chairholder (2 nd mandate) of the Chaire André Lussier de rhumatologie and is supported by the Canadian Institutes of Health Research (452873). Authors' contributions: HAC secured the funding for the project. BM and EB processed serum samples for mass spectrometry. DL analyzed the samples on the mass spectrometer and performed peptide/protein identification. DL and FMB contributed to mass spectrometry methodology. NC, SR, JM, GB and HAC were involved in patient enrollment and biobanking. BM and BR performed data analysis. NC contributed to statistical methodology and biobank coordination. BM, NC, BR, SR, JM, GB and HAC revised the manuscript. BM and HAC wrote the original draft and prepared the reviewed manuscript for submission. Acknowledgements: We are grateful for the contribution of rheumatologists Drs Guylaine Arsenault, Philippe Bilodeau, Lyne Bissonnette, Alessandra Bruns, Pierre Dagenais, Patrick Liang and Ariel Masetto to EUPA patient recruitment and follow-up. We extend our gratitude to research assistants Chantal Guillet, Noémie Poirier and Christine Rosa involved in sample biobanking of EUPA patients and at the Banque de Pathologies et Perturbations Immunes et Inflammatoires (BPPII) biobank, located at the Rheumatology Clinic of the CIUSSS de l’Estrie-CHUS in Canada. We express our sincere thanks to all the patients who participated in the EUPA study. References: Viatte S, Plant D, Bowes J, Lunt M, Eyre S, Barton A, Worthington J. Genetic markers of rheumatoid arthritis susceptibility in anti-citrullinated peptide antibody negative patients. Ann Rheum Dis. 2012;71(12):1984–90. https://doi.org/10.1136/annrheumdis-2011-201225 . Wu X, Liu Y, Jin S, Wang M, Jiao Y, Yang B, Lu X, Ji X, Fei Y, Yang H, Zhao L, Chen H, Zhang Y, Li H, Lipsky PE, Tsokos GC, Bai F, Zhang X. Single-cell sequencing of immune cells from anticitrullinated peptide antibody positive and negative rheumatoid arthritis. Nat Commun. 2021;12(1):4977. https://doi.org/10.1038/s41467-021-25246-7 . Barra L, Pope JE, Orav JE, Boire G, Haraoui B, Hitchon C, Keystone EC, Thorne JC, Tin D, Bykerk VP, the CATCH Investigators. Prognosis of Seronegative Patients in a Large Prospective Cohort of Patients with Early Inflammatory Arthritis. J Rhuematol. 2014;41(12):2361–9. https://doi.org/10.3899/jrheum.140082 . Meyer O. Anticitrullinated protein/peptide antibody assays in early rheumatoid arthritis for predicting five year radiographic damage. Ann Rheum Dis. 2003;62(2):120–6. https://doi.org/10.1136/ard.62.2.120 . Farragher TM, Lunt M, Fu B, Bunn D, Symmons DPM. Early treatment with, and time receiving, first disease-modifying antirheumatic drug predicts long-term function in patients with inflammatory polyarthritis. Ann Rheum Dis. 2010;69(4):689–95. https://doi.org/10.1136/ard.2009.108639 . Singh JA, Saag KG, Bridges SL, Akl EA, Bannuru RR, Sullivan MC, Vaysbrot E, McNaughton C, Osani M, Shmerling RH, Curtis JR, Furst DE, Parks D, Kavanaugh A, O’Dell J, King C, Leong A, Matteson EL, Schousboe JT, McAlindon T. 2015 American College of Rheumatology Guideline for the Treatment of Rheumatoid Arthritis. Arthritis Care Res. 2016;68(1):1–25. https://doi.org/10.1002/acr.22783 . Fraenkel L, Bathon JM, England BR, St.Clair EW, Arayssi T, Carandang K, Deane KD, Genovese M, Huston KK, Kerr G, Kremer J, Nakamura MC, Russell LA, Singh JA, Smith BJ, Sparks JA, Venkatachalam S, Weinblatt ME, Al-Gibbawi M, Akl EA. 2021 American College of Rheumatology Guideline for the Treatment of Rheumatoid Arthritis. Arthritis Rheumatol. 2021;73(7):1108–23. https://doi.org/10.1002/art.41752 . England BR, Tiong BK, Bergman MJ, Curtis JR, Kazi S, Mikuls TR, O’Dell JR, Ranganath VK, Limanni A, Suter LG, Michaud K. 2019 Update of the American College of Rheumatology Recommended Rheumatoid Arthritis Disease Activity Measures. Arthritis Care Res. 2019;71(12):1540–55. https://doi.org/10.1002/acr.24042 . Hambardzumyan K, Bolce RJ, Saevarsdottir S, Forslind K, Wallman JK, Cruickshank SE, Sasso EH, Chernoff D, van Vollenhoven RF. Association of a multibiomarker disease activity score at multiple time-points with radiographic progression in rheumatoid arthritis: Results from the SWEFOT trial. RMD Open. 2016;2(1):e000197. https://doi.org/10.1136/rmdopen-2015-000197 . Molenaar ETH, Voskuyl AE, Dinant HJ, Bezemer PD, Boers M, Dijkmans BAC. Progression of radiologic damage in patients with rheumatoid arthritis in clinical remission. Arthr Rhuem. 2004;50(1):36–42. https://doi.org/10.1002/art.11481 . Cohen G, Gossec L, Dougados M, Cantagrel A, Goupille P, Daures J, Rincheval N, Combe B. Radiological damage in patients with rheumatoid arthritis on sustained remission. Ann Rheum Dis. 2007;66(3):358–63. https://doi.org/10.1136/ard.2006.057497 . Gandjbakhch F, Conaghan PG, Ejbjerg B, Haavardsholm EA, Foltz V, Brown AK, Møller Døhn U, Lassere M, Freeston J, Bøyesen P, Bird P, Fautrel B, Hetland ML, Emery P, Bourgeois P, Hørslev-Petersen K, Kvien TK, McQUEEN F, Østergaard M. Synovitis and Osteitis Are Very Frequent in Rheumatoid Arthritis Clinical Remission: Results from an MRI Study of 294 Patients in Clinical Remission or Low Disease Activity State. J Rhuematol. 2011;38(9):2039–44. https://doi.org/10.3899/jrheum.110421 . Capelusnik D, Aletaha D. Baseline predictors of different types of treatment success in rheumatoid arthritis. Ann Rheum Dis. 2022;81(2):153–8. https://doi.org/10.1136/annrheumdis-2021-220853 . Luedders BA, Johnson TM, Sayles H, Thiele GM, Mikuls TR, O’Dell JR, England BR. Predictive ability, validity, and responsiveness of the multi-biomarker disease activity score in patients with rheumatoid arthritis initiating methotrexate. Semin Arthritis Rheum. 2020;50(5):1058–63. https://doi.org/10.1016/j.semarthrit.2020.05.019 . Vanier A, Smolen JS, Allaart CF, Van Vollenhoven R, Verschueren P, Vastesaeger N, Saevarsdottir S, Visser K, Aletaha D, Combe B, Fautrel B. An updated matrix to predict rapid radiographic progression of early rheumatoid arthritis patients: Pooled analyses from several databases. Rheumatology. 2020;59(8):1842–52. https://doi.org/10.1093/rheumatology/kez542 . Centola M, Cavet G, Shen Y, Ramanujan S, Knowlton N, Swan KA, Turner M, Sutton C, Smith DR, Haney DJ, Chernoff D, Hesterberg LK, Carulli JP, Taylor PC, Shadick NA, Weinblatt ME, Curtis JR. Development of a Multi-Biomarker Disease Activity Test for Rheumatoid Arthritis. PLoS ONE. 2013;8(4):e60635. https://doi.org/10.1371/journal.pone.0060635 . Bakker MF, Cavet G, Jacobs JW, Bijlsma JWJ, Haney DJ, Shen Y, Hesterberg LK, Smith DR, Centola M, van Roon JAG, Lafeber FPJG, Welsing PMJ. Performance of a multi-biomarker score measuring rheumatoid arthritis disease activity in the CAMERA tight control study. Ann Rheum Dis. 2012;71(10):1692–7. https://doi.org/10.1136/annrheumdis-2011-200963 . Curtis JR, Weinblatt ME, Shadick NA, Brahe CH, Østergaard M, Hetland ML, Saevarsdottir S, Horton M, Mabey B, Flake DD, Ben-Shachar R, Sasso EH, Huizinga TW. Validation of the adjusted multi-biomarker disease activity score as a prognostic test for radiographic progression in rheumatoid arthritis: A combined analysis of multiple studies. Arthritis Res Therapy. 2021;23(1):1. https://doi.org/10.1186/s13075-020-02389-4 . Curtis JR, Greenberg JD, Harrold LR, Kremer JM, Palmer JL. Influence of obesity, age, and comorbidities on the multi-biomarker disease activity test in rheumatoid arthritis. Semin Arthritis Rheum. 2018;47(4):472–7. https://doi.org/10.1016/j.semarthrit.2017.07.010 . Curtis JR, Flake DD, Weinblatt ME, Shadick NA, Østergaard M, Hetland ML, Brahe CH, Hwang YG, Furst DE, Strand V, Etzel CJ, Pappas DA, Wang X, Hwang CC, Sasso EH, Gutin A, Hitraya E, Lanchbury JS. Adjustment of the multi-biomarker disease activity score to account for age, sex and adiposity in patients with rheumatoid arthritis. Rheumatology. 2019;58(5):874–83. https://doi.org/10.1093/rheumatology/key367 . Meznerics FA, Kemény LV, Gunther E, Bakó E, Dembrovszky F, Szabó B, Ascsillán A, Lutz E, Csupor D, Hegyi P, Bánvölgyi A, Nagy G. Multibiomarker disease activity score: An objective tool for monitoring rheumatoid arthritis? A systematic review and meta-analysis. Rheumatology. 2023;62(6):2048–59. https://doi.org/10.1093/rheumatology/keac715 . Carrasco-Zanini J, Pietzner M, Davitte J, Surendran P, Croteau-Chonka DC, Robins C, Torralbo A, Tomlinson C, Grünschläger F, Fitzpatrick N, Ytsma C, Kanno T, Gade S, Freitag D, Ziebell F, Haas S, Denaxas S, Betts JC, Wareham NJ, Langenberg C. Proteomic signatures improve risk prediction for common and rare diseases. Nat Med. 2024;30(9):2489–98. https://doi.org/10.1038/s41591-024-03142-z . Jin L, Wang F, Wang X, Harvey BP, Bi Y, Hu C, Cui B, Darcy AT, Maull JW, Phillips BR, Kim Y, Jenkins GJ, Sornasse TR, Tian Y. Identification of Plasma Biomarkers from Rheumatoid Arthritis Patients Using an Optimized Sequential Window Acquisition of All THeoretical Mass Spectra (SWATH) Proteomics Workflow. Proteomes. 2023;11(4):32. https://doi.org/10.3390/proteomes11040032 . Hu C, Dai Z, Xu J, Zhao L, Xu Y, Li M, Yu J, Zhang L, Deng H, Liu L, Zhang M, Huang J, Wu L, Chen G. Proteome Profiling Identifies Serum Biomarkers in Rheumatoid Arthritis. Front Immunol. 2022;13:865425. https://doi.org/10.3389/fimmu.2022.865425 . Mun S, Lee J, Park M, Shin J, Lim M-K, Kang H-G. Serum biomarker panel for the diagnosis of rheumatoid arthritis. Arthritis Res Therapy. 2021;23(1):31. https://doi.org/10.1186/s13075-020-02405-7 . O’Neil LJ, Spicer V, Smolik I, Meng X, Goel RR, Anaparti V, Wilkins J, El-Gabalawy HS. Association of a Serum Protein Signature With Rheumatoid Arthritis Development. Arthritis Rheumatol. 2021;73(1):78–88. https://doi.org/10.1002/art.41483 . O’Neil LJ, Meng X, Mcfadyen C, Fritzler MJ, El-Gabalawy HS. Serum proteomic networks associate with pre-clinical rheumatoid arthritis autoantibodies and longitudinal outcomes. Front Immunol. 2022;13:958145. https://doi.org/10.3389/fimmu.2022.958145 . Carrier N, De Brum-Fernandes AJ, Liang P, Masetto A, Roux S, Biln NK, Maksymowych WP, Boire G. Impending radiographic erosive progression over the following year in a cohort of consecutive patients with inflammatory polyarthritis: Prediction by serum biomarkers. RMD Open. 2020;6(1):e001191. https://doi.org/10.1136/rmdopen-2020-001191 . Nguyen HD, Lortie A, Mbous Nguimbus L, Marrugo J, Allard-Chamard H, Bouchard L, Boire G, Scott MS, Roux S. Osteoclast microRNA Profiling in Rheumatoid Arthritis to Capture the Erosive Factor. JBMR Plus. 2023;7(8):e10776. https://doi.org/10.1002/jbm4.10776 . Kitata RB, Yang J, Chen Y. Advances in data-independent acquisition mass spectrometry towards comprehensive digital proteome landscape. Mass Spectrom Rev. 2023;42(6):2324–48. https://doi.org/10.1002/mas.21781 . Zhao Y, Xue Q, Wang M, Meng B, Jiang Y, Zhai R, Zhang Y, Dai X, Fang X. Evolution of Mass Spectrometry Instruments and Techniques for Blood Proteomics. J Proteome Res. 2023;22(4):1009–23. https://doi.org/10.1021/acs.jproteome.3c00102 . Boire G, Cossette P, de Brum-Fernandes AJ, Liang P, Niyonsenga T, Zhou ZJ, Carrier N, Daniel C, Ménard H-A. Anti-Sa antibodies and antibodies against cyclic citrullinated peptide are not equivalent as predictors of severe outcomes in patients with recent-onset polyarthritis. Arthritis Res Therapy. 2005;7(3):R592–603. https://doi.org/10.1186/ar1719 . Carrier N, Roux S, Masetto A, De Brum-Fernandes AJ, Liang P, Maoui M, Boire G. Changes at Presentation in Patients With Early Rheumatoid Arthritis: A 24-Year Study of the Early Undifferentiated Polyarthritis (EUPA) Cohort. J Rhuematol. 2024;52(2). https://doi.org/10.3899/jrheum.2024-0560 . jrheum.2024 – 0560. van der Heijde D. How to read radiographs according to the Sharp/van der Heijde method. J Rhuematol. 2000;27(1):261–3. Carrier N, Marotta A, de Brum-Fernandes AJ, Liang P, Masetto A, Ménard HA, Maksymowych WP, Boire G. Serum levels of 14-3-3η protein supplement C-reactive protein and rheumatoid arthritis-associated antibodies to predict clinical and radiographic outcomes in a prospective cohort of patients with recent-onset inflammatory polyarthritis. Arthritis Res Therapy. 2016;18(1):37. https://doi.org/10.1186/s13075-016-0935-z . Demichev V, Messner CB, Vernardis SI, Lilley KS, Ralser M. DIA-NN: Neural networks and interference correction enable deep proteome coverage in high throughput. Nat Methods. 2020;17(1):41–4. https://doi.org/10.1038/s41592-019-0638-x . Cox J, Hein MY, Luber CA, Paron I, Nagaraj N, Mann M. Accurate Proteome-wide Label-free Quantification by Delayed Normalization and Maximal Peptide Ratio Extraction, Termed MaxLFQ. Mol Cell Proteom. 2014;13(9):2513–26. https://doi.org/10.1074/mcp.M113.031591 . Wieczorek S, Combes F, Lazar C, Giai Gianetto Q, Gatto L, Dorffer A, Hesse A-M, Couté Y, Ferro M, Bruley C, Burger T. DAPAR & ProStaR: Software to perform statistical analyses in quantitative discovery proteomics. Bioinformatics. 2017;33(1):135–6. https://doi.org/10.1093/bioinformatics/btw580 . R Core Team. (2024). R: A Language and Environment for Statistical Computing [Computer software]. R Foundation for Statistical Computing. https://www.R-project.org Lê S, Josse J, Husson F. FactoMineR: An R Package for Multivariate Analysis. J Stat Softw. 2008;25(1). https://doi.org/10.18637/jss.v025.i01 . Kassambara A, Mundt F. (2016). factoextra: Extract and Visualize the Results of Multivariate Data Analyses (p. 1.0.7) [Dataset]. https://doi.org/10.32614/CRAN.package.factoextra Halekoh U, Højsgaard S, Yan J. The R Package geepack for Generalized Estimating Equations. J Stat Softw. 2006;15(2). https://doi.org/10.18637/jss.v015.i02 . Lüdecke D, Makowski D, Ben-Shachar MS, Patil I, Wiernik BM, Bacher E, Thériault R. (2022). easystats: Framework for Easy Statistical Modeling, Visualization, and Reporting (p. 0.7.4) [Dataset]. https://doi.org/10.32614/CRAN.package.easystats Robin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez J-C, Müller M. pROC: An open-source package for R and S + to analyze and compare ROC curves. BMC Bioinformatics. 2011;12(1):77. https://doi.org/10.1186/1471-2105-12-77 . Ulgen E, Ozisik O, Sezerman OU. pathfindR: An R Package for Comprehensive Identification of Enriched Pathways in Omics Data Through Active Subnetworks. Front Genet. 2019;10:858. https://doi.org/10.3389/fgene.2019.00858 . Kanehisa M, Furumichi M, Sato Y, Matsuura Y, Ishiguro-Watanabe M. KEGG: Biological systems database as a model of the real world. Nucleic Acids Res. 2025;53(D1):D672–7. https://doi.org/10.1093/nar/gkae909 . Bruynesteyn K, Van Der Heijde D, Boers M, Saudan A, Peloso P, Paulus H, Houben H, Griffiths B, Edmonds J, Bresnihan B, Boonen A, Van Der Linden S. Determination of the minimal clinically important difference in rheumatoid arthritis joint damage of the Sharp/van der Heijde and Larsen/Scott scoring methods by clinical experts and comparison with the smallest detectable difference. Arthr Rhuem. 2002;46(4):913–20. https://doi.org/10.1002/art.10190 . Takeuchi T, Soen S, Ishiguro N, Yamanaka H, Tanaka S, Kobayashi M, Okubo N, Nitta T, Tanaka Y. Predictors of new bone erosion in rheumatoid arthritis patients receiving conventional synthetic disease-modifying antirheumatic drugs: Analysis of data from the DRIVE and DESIRABLE studies. Mod Rheumatol. 2021;31(1):34–41. https://doi.org/10.1080/14397595.2019.1703484 . O’Neil LJ, Hu P, Liu Q, Islam MM, Spicer V, Rech J, Hueber A, Anaparti V, Smolik I, El-Gabalawy HS, Schett G, Wilkins JA. Proteomic Approaches to Defining Remission and the Risk of Relapse in Rheumatoid Arthritis. Front Immunol. 2021;12:729681. https://doi.org/10.3389/fimmu.2021.729681 . He Z, Glass MC, Venkatesan P, Feser ML, Lazaro L, Okada LY, Tran NTT, He YD, Zaim R, Bennett S, Ravisankar CE, Dornisch P, Arishi EM, Asamoah NA, Barzideh AG, Becker S, Bemis LA, Buckner EA, Collora JH, Gillespie CE. M. A. (2024). Systemic inflammation and lymphocyte activation precede rheumatoid arthritis . https://doi.org/10.1101/2024.10.25.620344 Cho A, Ahn J, Kim A, Lee YJ, Song YW, Tanaka Y, Yi EC. A multi-biomarker panel for predicting Tocilizumab response in Rheumatoid arthritis patients. Translational Res. 2024;273:23–31. https://doi.org/10.1016/j.trsl.2024.07.001 . Jiang Y, Hu Z, Huang R, Ho K, Wang P, Kang J. Metabolic reprogramming and macrophage expansion define ACPA-negative rheumatoid arthritis: Insights from single-cell RNA sequencing. Front Immunol. 2025;15:1512483. https://doi.org/10.3389/fimmu.2024.1512483 . Banda NK, Mehta G, Ferreira VP, Cortes C, Pickering MC, Pangburn MK, Arend WP, Holers VM. Essential Role of Surface-Bound Complement Factor H in Controlling Immune Complex–Induced Arthritis. J Immunol. 2013;190(7):3560–9. https://doi.org/10.4049/jimmunol.1203271 . Hietala MA, Jonsson I-M, Tarkowski A, Kleinau S, Pekna M. Complement Deficiency Ameliorates Collagen-Induced Arthritis in Mice. J Immunol. 2002;169(1):454–9. https://doi.org/10.4049/jimmunol.169.1.454 . Hietala MA, Nandakumar KS, Persson L, Fahlén S, Holmdahl R, Pekna M. Complement activation by both classical and alternative pathways is critical for the effector phase of arthritis. Eur J Immunol. 2004;34(4):1208–16. https://doi.org/10.1002/eji.200424895 . Banda NK, Levitt B, Wood AK, Takahashi K, Stahl GL, Holers VM, Arend WP. Complement activation pathways in murine immune complex-induced arthritis and in C3a and C5a generation in vitro . Clin Exp Immunol. 2009;159(1):100–8. https://doi.org/10.1111/j.1365-2249.2009.04035.x . Banda NK, Takahashi M, Levitt B, Glogowska M, Nicholas J, Takahashi K, Stahl GL, Fujita T, Arend WP, Holers VM. Essential Role of Complement Mannose-Binding Lectin-Associated Serine Proteases-1/3 in the Murine Collagen Antibody-Induced Model of Inflammatory Arthritis. J Immunol. 2010;185(9):5598–606. https://doi.org/10.4049/jimmunol.1001564 . Sato Y, Sato R, Watanabe H, Kogure A, Watanabe K, Nishimaki T, Kasukawa R, Kuraya M, Fujita T. Complement activating properties of monoreactive and polyreactive IgM rheumatoid factors. Ann Rheum Dis. 1993;52(11):795–800. https://doi.org/10.1136/ard.52.11.795 . Sabharwal UK, Vaughan JH, Fong S, Bennett PH, Carson DA, Curd JG. Activation of the classical pathway of complement by rheumatoid factors. Arthr Rhuem. 1982;25(2):161–7. https://doi.org/10.1002/art.1780250208 . Jahn B, Von Kempis J, Krämer KL, Filsinger S, Hänsch GM. Interaction of the terminal complement components C5b-9 with synovial fibroblasts: Binding to the membrane surface leads to increased levels in collagenase-specific mRNA. Immunology. 1993;78(2):329–34. Crockard AD, Thompson JM, McBride SJ, Edgar JD, McNeill TA, Bell AL. Markers of inflammatory activation: Upregulation of complement receptors CR1 and CR3 on synovial fluid neutrophils from patients with inflammatory joint disease. Clin Immunol Immunopathol. 1992;65(2):135–42. https://doi.org/10.1016/0090-1229(92)90216-B . Merkel PA, George MD, Yue H, Popov T, Kronbichler A, Little MA, Jayne DRW. Safety of Avacopan for the Treatment of Antineutrophil Cytoplasmic Antibody–Associated Vasculitis: Combined Data From Three Clinical Trials. ACR Open Rheumatol. 2025;7(4):e70001. https://doi.org/10.1002/acr2.70001 . Trouw LA, Pickering MC, Blom AM. The complement system as a potential therapeutic target in rheumatic disease. Nat Rev Rheumatol. 2017;13(9):538–47. https://doi.org/10.1038/nrrheum.2017.125 . Storey JD. A Direct Approach to False Discovery Rates. J Royal Stat Soc Ser B: Stat Methodol. 2002;64(3):479–98. https://doi.org/10.1111/1467-9868.00346 . Pope JE, Choy EH. C-reactive protein and implications in rheumatoid arthritis and associated comorbidities. Semin Arthritis Rheum. 2021;51(1):219–29. https://doi.org/10.1016/j.semarthrit.2020.11.005 . Chambers RE, MacFarlane DG, Whicher JT, Dieppe PA. Serum amyloid-A protein concentration in rheumatoid arthritis and its role in monitoring disease activity. Ann Rheum Dis. 1983;42(6):665–7. https://doi.org/10.1136/ard.42.6.665 . Connolly M, Mullan RH, Mccormick J, Matthews C, Sullivan O, Kennedy A, FitzGerald O, Poole AR, Bresnihan B, Veale DJ, Fearon U. Acute-phase serum amyloid A regulates tumor necrosis factor α and matrix turnover and predicts disease progression in patients with inflammatory arthritis before and after biologic therapy. Arthr Rhuem. 2012;64(4):1035–45. https://doi.org/10.1002/art.33455 . Sorić Hosman I, Kos I, Lamot L. Serum Amyloid A in Inflammatory Rheumatic Diseases: A Compendious Review of a Renowned Biomarker. Front Immunol. 2021;11:631299. https://doi.org/10.3389/fimmu.2020.631299 . Ma MHY, Defranoux N, Li W, Sasso EH, Ibrahim F, Scott DL, Cope AP. A multi-biomarker disease activity score can predict sustained remission in rheumatoid arthritis. Arthritis Res Therapy. 2020;22(1):158. https://doi.org/10.1186/s13075-020-02240-w . Naka T, Fujimoto M. LRG is a novel inflammatory marker clinically useful for the evaluation of disease activity in rheumatoid arthritis and inflammatory bowel disease. Immunological Med. 2018;41(2):62–7. https://doi.org/10.1080/13497413.2018.1481582 . Zou L, Fan Q, Liu Y, He H, Jia C. Leucine-rich alpha-2-glycoprotein 1 (LRG1) decrement during biologics therapy and its correlation with disease features and treatment outcomes in rheumatoid arthritis patients. Clin Rheumatol. 2025;44(9):3391–8. https://doi.org/10.1007/s10067-025-07545-2 . Qin Q, Song R, Du P, Gao C, Yao Q, Zhang J-A. (2021). Systemic Proteomic Analysis Reveals Distinct Exosomal Protein Profiles in Rheumatoid Arthritis. Journal of Immunology Research , 2021 , 9421720. https://doi.org/10.1155/2021/9421720 Wang Q, Feng D, Song Y, Hu Z, Lu Q, Zhao M. Biased Usage of V/D/J Genes and Clonal Diversity in IgG Repertoires Correlates with Disease Activity and Clinical Features in Systemic Autoimmune Diseases. Immunol Investig. 2025;54(8):1461–81. https://doi.org/10.1080/08820139.2025.2550374 . Lee PY, Osman J, Low TY, Jamal R. Plasma/Serum Proteomics: Depletion Strategies for Reducing High-Abundance Proteins for Biomarker Discovery. Bioanalysis. 2019;11(19):1799–812. https://doi.org/10.4155/bio-2019-0145 . Tables: Table 1: Clinical and demographic description of patients from the study cohort. Seronegative Seropositive Demographic Features Individuals, n 48 59 Female, % (n) 69% (33) 59% (35) Age, years 63.9 (54.1–72.3) 59.5 (51.5–67.5) Ethnicity - White, % (n) 100% (48) 97% (57) Non-Smoker, % (n) 48% (23) 36% (21) a Smoker, % (n) 4% (2) 17% (10) a Former-Smoker, % (n) 48% (23) 46% (27) a Baseline Clinical Features Body Mass Index, kg/m² 27.4 (24.4–31.2) 27.1 (23.6–30.1) RF positive, % (n) 0% (0) 3% (2) ACPA positive, % (n) 0% (0) 19% (11) ACPA/RF positive, % (n) 0% (0) 78% (46) Symptom duration, months 3.4 (1.9-6) b 4.1 (2.8–6.6) Rheumatic disease comorbidity index (RDCI) 1 (0–2) 1 (0–2) CRP, mg/L 11.8 (4.3–27.4) 11.6 (4.1–30.5) ESR, mm/h 21 (13–36) b 30.5 (20-48.8) a Baseline Disease Activity Measures Swollen joint count (68) 10 (8–14) 13 (8–21) Tender joint count (66) 12 (6–18) 15 (10–20) DAS28-CRP 5.3 (4.1-6) c 5.2 (4.1–6.2) a SDAI 31.9 (23.9–40.7) d 34.6 (22.6–45.3) CDAI 30.8 (21.8–38.6) d 33.8 (20.4–42.9) HAQ 0.9 (0.2–1.1) c 1.1 (0.6–1.5) e PGA 62 (49–84) c 57.5 (34.5–77.2) EGA 5 (2.9–6.7) b 5.8 (3.5–7.2) Treatment and Outcomes DMARD exposure, months 7 (5.2–8.8) 6.8 (4.9–8.5) Prednisone use within 12m, % (n) 17% (8) 12% (7) MTX alone use within 12m, % (n) 33% (16) 34% (20) MTX combination use within 12m, % (n) 40% (19) 53% (31) bDMARD use within 12m, % (n) 10% (5) 5% (3) Patient in remission at 12m, % (n) 48% (23) 54% (32) Erosive patients at inclusion, % (n) 10% (5) b 17% (10) Erosive patients at 30m, % (n) 15% (7) b 37% (22) SvH > = 5 patients, % (n) 42% (20) b 61% (36) Footnote: Variables are presented as either median (IQR; 25th-75th) or % (n) without imputation of missing values. The MTX combination category comprises patients treated with MTX in combination with at least one additional DMARD at inclusion, or who escalated to an additional DMARD within 12 months. (a: n = 58; b: n = 47; c: ; n = 45; d: n = 44; e: n = 56) Anti-Cyclic Citrullinated Peptide Antibody (ACPA), Rheumatoid Factor (RF), C-Reactive Protein (CRP), Erythrocyte Sedimentation Rate (ESR), Disease Activity Score in 28 joints with CRP (DAS28-CRP), Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), Health Assessment Questionnaire (HAQ), Patient Global Assessment (PGA), Evaluator Global Assessment (EGA), Sharp-van der Heijde total score (SvH). Table 2: Patient characteristics by clusters based on baseline vs. 12-month proteome. Baseline Post-treatment B1 B2 P value P1 P2 P value Demographic Features Individuals, n 49 58 NA 28 79 NA Gender - Females, % (n) 57% (28) 69% (40) 0.287 50% (14) 68% (54) 0.132 Age, years 59 (48.5–67.5) 62 (53.6–69.5) 0.328 61.2 (53.4–69.5) 59.5 (51.5–68.5) 0.840 Body Mass Index, kg/m² 27.3 (24.4–29.3) 27.2 (23.4–31.4) 0.571 27.1 (23.6–30.4) 27.4 (23.8–31) 0.642 Baseline Clinical Features ACPA or RF positive, % (n) 69% (34) * 43% (25) * 0.011 61% (17) 53% (42) 0.639 Symptom duration, months 4.1 (2.3–6.5) 3.6 (2.1-6) a 0.434 4.8 (2.6–6.6) 3.9 (2.1–6.1) b 0.211 Swollen joint count (68) 11 (7–18) 12 (8–20) 0.389 10 (7–17) 12 (8–19) 0.558 Tender joint count (66) 13 (7–18) 15 (7–20) 0.386 14 (8–17) 14 (7–19) 0.739 CRP, mg/L 11.7 (5.6–27.1) 11.9 (3.6–29.5) 0.970 10.2 (3.1–27.5) 12.7 (5.5–28.6) 0.412 ESR, mm/h 30 (18–42) 23 (16-46.5) c 0.414 30 (16-46.5) 24 (16–42) d 0.980 Baseline Disease Activity Measures DAS28-CRP 5 (4–6) 5.5 (4.6–6.2) e 0.211 5.1 (4.1–6.2) 5.3 (4.2–6.1) f 0.692 SDAI 30.2 (22.5–41.9) 37.8 (24.5–46.1) g 0.190 30.2 (22.4–44.2) 34.4 (23-44.9) h 0.725 CDAI 28.4 (19.9–38.6) 33.9 (22–40) g 0.198 29.9 (22.3–41.5) 32.6 (20.9–39) h 0.964 HAQ 1 (0.6–1.3) 0.9 (0.4–1.5) g 0.881 0.9 (0.5–1.4) 1 (0.5–1.4) i 0.912 12-month Clinical Features Swollen joint count (68) 2 (0–8) 2 (0–6) 0.610 2 (0–8) 2 (0–6) 0.370 Tender joint count (66) 2 (0–6) 2 (0–6) 0.814 2 (0–6) 2 (0–6) 0.539 CRP, mg/L 1 (1–8) 1 (1-6.4) 0.765 1 (1–4) 1 (1-8.8) 0.110 ESR, mm/h 16 (6–20) 14 (10–22) 0.972 10.5 (8-18.5) 16 (9.5–22) 0.398 12-month Disease Activity Measures DAS28-CRP 3.2 (1.7-4) 2.5 (1.9–3.7) 0.819 2.5 (1.9–3.8) 2.6 (1.8-4) 0.935 SDAI 10.1 (3.5–18.7) 8.1 (4–14) a 0.626 9.4 (4.2–19.7) 9.5 (3.6–15) b 0.606 CDAI 9.8 (3.4–17.9) 7.8 (3.7–13.9) a 0.660 8.9 (4.1–19.2) 8.6 (3.4–14.5) b 0.469 HAQ 0.1 (0-0.5) 0.2 (0-0.8) 0.248 0.2 (0-0.6) 0.2 (0-0.6) 0.647 Treatment and Outcomes DMARD exposure, months 7 (5.3–8.5) 6.8 (5.2–8.9) 0.802 6.8 (4.9–8.5) 7.2 (5.2–8.6) 0.530 Prednisone use within 12m, % (n) 8% (4) 19% (11) 0.162 7% (2) 16% (13) 0.344 MTX alone use within 12m, % (n) 39% (19) 29% (17) 0.408 32% (9) 34% (27) 1.000 MTX combination use within 12m, % (n) 47% (23) 47% (27) 1.000 46% (13) 47% (37) 1.000 bDMARD use within 12m, % (n) 4% (2) 10% (6) 0.288 0% (0) 10% (8) b 0.107 Patient in remission at 12m, % (n) 49% (24) 53% (31) 0.790 54% (15) 51% (40) 0.962 Erosive patients at 30m, % (n) 29% (14) 26% (15) a 0.967 21% (6) 29% (23) b 0.566 SvH > = 5 patients, % (n) 57% (28) 48% (28) a 0.529 50% (14) 53% (42) b 0.897 Footnote: Variables are presented as either median (IQR; 25th-75th) or % (n) without imputation of missing values. The Mann-Whitney U test was used for continuous variables, and Chi-square or Fisher’s exact tests were applied for categorical variables. The MTX combination category comprises patients treated with MTX in combination with at least one additional DMARD at inclusion, or who escalated to an additional DMARD within 12 months. * Indicate statistical significance at p ≤0.05. (a: n = 57; b: n = 78; c: n = 56; d: n = 77; e: n = 54; f: n = 75; g: n = 53; h: n = 74; i: n = 73) Anti-Cyclic Citrullinated Peptide Antibody (ACPA), Rheumatoid Factor (RF), C-Reactive Protein (CRP), Erythrocyte Sedimentation Rate (ESR), Disease Activity Score in 28 joints with CRP (DAS28-CRP), Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), Health Assessment Questionnaire (HAQ). Additional Declarations Competing interest reported. GB declares an advisory relationship with Otsuka Canada, and received honoraria for presentations and unrestricted research funding from Biocon Biologics Canada Inc. SR declares an advisory relationship with Amgen Canada, Kyowa Kirin, and Apotex. SR received speaker fees and funding from Kyowa Kirin, as well as funding from Insmed. HAC declares an advisory relationship with Abbvie, Amgen Canada, AstraZeneca, Celltrion, Eli Lilly & Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Novartis Canada, Pfizer Canada, Sandoz Canada and Sobi. HAC received speaker fees from Abbvie, Amgen Canada, AstraZeneca, Bristol Myers Squibb, Celltrion, Eli Lilly & Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Mantra Pharma, Novartis Pharmaceuticals Canada, Pfizer Canada and Sobi. HAC was awarded funding from Abbvie, AstraZeneca, Bristol Myers Squibb, Daiichi Sankyo, Eli Lilly & Co, Fresenius Kabi USA, Neomed, Novartis Pharmaceuticals Canada, Pfizer Canada, Sanofi and Viela Bio. All other authors declare no known competing interests related to the work reported in this article. Supplementary Files AdditionalFile1SupplementaryFigures19.docx AdditionalFile2SuppTable1.xlsx AdditionalFile3SuppTable2.xlsx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 15 May, 2026 Reviewers agreed at journal 13 May, 2026 Reviewers invited by journal 01 May, 2026 Editor assigned by journal 10 Apr, 2026 Submission checks completed at journal 10 Apr, 2026 First submitted to journal 08 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9359440","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":637476767,"identity":"30ee51b0-577a-4c3d-84cf-dd59f5cd4246","order_by":0,"name":"Benoît Marchand","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Benoît","middleName":"","lastName":"Marchand","suffix":""},{"id":637476768,"identity":"83826358-222f-4dca-b413-cc26c7e390c7","order_by":1,"name":"Nathalie Carrier","email":"","orcid":"","institution":"Centre Hospitalier Universitaire de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Nathalie","middleName":"","lastName":"Carrier","suffix":""},{"id":637476769,"identity":"656719d8-d54f-44d4-ab80-8fa4ce9ad0ca","order_by":2,"name":"Elizabeth Beaulieu","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Elizabeth","middleName":"","lastName":"Beaulieu","suffix":""},{"id":637476770,"identity":"c0fb5235-40a1-4835-93fd-77b110ff0178","order_by":3,"name":"Dominique Lévesque","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Dominique","middleName":"","lastName":"Lévesque","suffix":""},{"id":637476771,"identity":"2c79d653-7cef-4c0d-8d06-af4623816d82","order_by":4,"name":"Barath Ramanathan","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Barath","middleName":"","lastName":"Ramanathan","suffix":""},{"id":637476772,"identity":"c1a4468a-e5e6-4c67-b230-6ff2cc9021b8","order_by":5,"name":"François-Michel Boisvert","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"François-Michel","middleName":"","lastName":"Boisvert","suffix":""},{"id":637476774,"identity":"03622c07-d9ff-469a-a243-354f091f2712","order_by":6,"name":"Sophie Roux","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Sophie","middleName":"","lastName":"Roux","suffix":""},{"id":637476776,"identity":"0c6b6c25-320e-49a6-840f-f620f263ea80","order_by":7,"name":"Javier Marrugo","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Javier","middleName":"","lastName":"Marrugo","suffix":""},{"id":637476778,"identity":"caf88fe7-ed17-47b4-bd1d-82a91c577cd3","order_by":8,"name":"Gilles Boire","email":"","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":false,"prefix":"","firstName":"Gilles","middleName":"","lastName":"Boire","suffix":""},{"id":637476779,"identity":"cde5e3d1-ba04-49bf-bfb4-767ac4d60bb2","order_by":9,"name":"Hugues Allard-Chamard","email":"data:image/png;base64,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","orcid":"","institution":"Université de Sherbrooke","correspondingAuthor":true,"prefix":"","firstName":"Hugues","middleName":"","lastName":"Allard-Chamard","suffix":""}],"badges":[],"createdAt":"2026-04-08 16:24:47","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9359440/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9359440/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":109075453,"identity":"3441b6a2-c83e-4748-acfa-d87c8cd8455c","added_by":"auto","created_at":"2026-05-12 10:56:19","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":732452,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRA patient clustering based on serum proteomic landscapes.\u003c/strong\u003e Hierarchical Clustering on Principal Component (HCPC) analyses were performed on RA patients’ serum proteome measured at \u003cstrong\u003eA)\u003c/strong\u003e baseline or \u003cstrong\u003eB)\u003c/strong\u003e 12-month follow-up visit. \u003cstrong\u003eC-D)\u003c/strong\u003e Z-scores were calculated for serum levels of complement proteins (KEGG pathway hsa04610) across all patients either at baseline or 12-month. The overall complement proteins levels of patients are shown by overlaying the average z-score as color gradient on PCA plots using \u003cstrong\u003eC)\u003c/strong\u003e baseline or \u003cstrong\u003eD)\u003c/strong\u003e 12-month proteomic data. \u003cstrong\u003eE)\u003c/strong\u003eSankey plot illustrating RA patients’ trajectory between baseline and 12-month clustering. \u003cstrong\u003eF)\u003c/strong\u003e The average z-scores for complement proteins serum levels of patients grouped according to cluster trajectories. Benjamini-Hochberg adjusted p ≤ 0.05 = * and ≤ 0.001 = ***.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/08e7cd83b77cfd32c03daf0c.png"},{"id":109075455,"identity":"728231f3-7250-4ecc-80c2-c09da5055fbc","added_by":"auto","created_at":"2026-05-12 10:56:19","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":986239,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRA disease activity-associated proteins identified by longitudinal proteomic analysis.\u003c/strong\u003e Generalized Estimating Equation (GEE) models adjusted for age, gender, and symptom duration were performed on longitudinal (baseline and 12-month) RA patients’ serum proteins levels for binary outcomes of Disease Activity (DA) at 12-month. Volcano plots of proteins associated with DA in \u003cstrong\u003eA)\u003c/strong\u003e seropositive (n=57) or \u003cstrong\u003eB)\u003c/strong\u003e seronegative (n=42) RA patients. The top 10 significant proteins are labelled and horizontal line show p threshold of 0.05. \u003cstrong\u003eC)\u003c/strong\u003e Volcano plot of DA-associated proteins identified using GEE models performed using full RA patient cohort, with serology as additional confounder (n=99; seropositive and seronegative combined). A total of eight patients were excluded from the longitudinal GEE analyses: three had their baseline DAS28 scores falling outside our predefined binomial DA outcome threshold, four lacked available baseline DAS28 values, and one was missing symptom duration information. \u003cstrong\u003eD)\u003c/strong\u003e Venn diagram showing DA-associated proteins identification overlap between GEE analyses performed relative to RA patients’ serology or on full cohort. \u003cstrong\u003eE)\u003c/strong\u003e KEGG pathway enrichment analysis was conducted using the DA-associated proteins identified through GEE models (p ≤ 0.05) performed using full patient cohort. Plot showing the top 10 KEGG enriched terms with at least 2 proteins and a Benjamini-Hochberg adjusted p value ≤0.05. Protein names separated by semicolons indicates that peptides identified by mass spectrometry were common to both.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/22d9f339649fbc98a267af5c.png"},{"id":109075853,"identity":"e1af8947-68f4-4954-a1e5-557ee25576f7","added_by":"auto","created_at":"2026-05-12 10:57:50","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":798479,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRA erosion-associated proteins identified by longitudinal serum proteomic analysis.\u003c/strong\u003e Generalized Estimating Equation (GEE) models adjusted for age, gender, and symptom duration were performed on longitudinal (baseline and 12-month) serum proteins levels for binary erosion outcome. Volcano plots of erosion-associated proteins identified using GEE in \u003cstrong\u003eA)\u003c/strong\u003eseropositive (n=59) or \u003cstrong\u003eB)\u003c/strong\u003e seronegative (n=46) RA patients. The top 10 proteins are labelled and horizontal line show p threshold of 0.05. \u003cstrong\u003eC)\u003c/strong\u003eVolcano plot showing erosion-associated proteins identified through GEE models performed on full RA patient cohort, with serology as additional confounder (n=105; seropositive and seronegative combined). Two patients were excluded from the longitudinal GEE analyses for erosion outcome due to missing Sharp score and symptom duration values. \u003cstrong\u003eD)\u003c/strong\u003e Venn diagram showing erosion-associated proteins identification overlap across GEE analyses conducted using data subgrouped by patients’ serology or on full cohort. \u003cstrong\u003eE)\u003c/strong\u003eKEGG pathway enrichment analysis was performed with the erosion-associated proteins identified by GEE models using full patient cohort. Plot showing the KEGG enriched terms with at least 2 proteins and a Benjamini-Hochberg adjusted p value ≤0.05. Protein names separated by semicolons indicates that peptides identified by mass spectrometry were common to both.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/f70231c7fdd82a534287891c.png"},{"id":109075854,"identity":"3d61d6b5-45c9-4f69-b116-c425791be574","added_by":"auto","created_at":"2026-05-12 10:57:50","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":301741,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of baseline serum proteins predictors of RA outcomes. \u003c/strong\u003eGeneralized Linear Models (GLM) analyses with binary outcomes (logistic regression) were performed using RA patients’ baseline serum proteins levels. The models were adjusted for age, gender, serology, and symptom duration. \u003cstrong\u003eA)\u003c/strong\u003e Forest plot showing the Odds Ratio (OR) and 95% CI of potential protein predictors of DA identified by GLM models with p values ≤0.05. One patient was excluded from this analysis due to missing information on symptom duration. \u003cstrong\u003eB)\u003c/strong\u003e Forest plot showing the OR and 95% CI of potential protein predictors of erosion identified by GLM models with p values ≤0.05. The p values are shown as a color gradient and proteins labelled in blue had Fold Changes (FC) above 1.25x or below 0.75x. Two patients were excluded from this analysis due to missing values for erosion (Sharp score) or symptom duration. Protein names separated by semicolons indicates that peptides identified by mass spectrometry were common to both.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/d0095e272b868dd5738d613d.png"},{"id":109075456,"identity":"4c1c0bca-908c-4ca5-b24e-0a6867c963d6","added_by":"auto","created_at":"2026-05-12 10:56:19","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":566527,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePerformance assessment of potential RA predictors of Disease Activity (DA) and Erosion. \u003c/strong\u003eROC curves plots showing sensitivity and specificity of \u003cstrong\u003eA) \u003c/strong\u003ethe univariate predictors of disease activity or \u003cstrong\u003eB)\u003c/strong\u003e the top 5 combinations of those predictors. Plots for erosion predictors ROC curves of \u003cstrong\u003eC)\u003c/strong\u003e the top 5 univariate proteins or \u003cstrong\u003eD) \u003c/strong\u003ethe top 5 combinations are shown. All GLM models used for calculating ROC curves were adjusted for age, serology, and symptom duration confounders. Protein names separated by semicolons indicates that peptides identified by mass spectrometry were common to both.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/1b896f84538d22aac342d0aa.png"},{"id":109081055,"identity":"07ed0ab6-64e9-4286-bd9c-8c0561ce606f","added_by":"auto","created_at":"2026-05-12 11:55:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3538965,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/6a2e4d12-cb24-4902-a3ce-403215a739b8.pdf"},{"id":109075452,"identity":"49b2752d-8e31-4c33-b05b-7aa0d957da59","added_by":"auto","created_at":"2026-05-12 10:56:18","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3030630,"visible":true,"origin":"","legend":"","description":"","filename":"AdditionalFile1SupplementaryFigures19.docx","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/86af6d4e759462be5c85b986.docx"},{"id":109075454,"identity":"3e079cd9-248d-49ab-947d-ef1482ed1b99","added_by":"auto","created_at":"2026-05-12 10:56:19","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":11638,"visible":true,"origin":"","legend":"","description":"","filename":"AdditionalFile2SuppTable1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/965a31e60943af0c244911da.xlsx"},{"id":109075851,"identity":"7c6e5889-80bd-411d-88bf-984ed565d511","added_by":"auto","created_at":"2026-05-12 10:57:49","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":16945,"visible":true,"origin":"","legend":"","description":"","filename":"AdditionalFile3SuppTable2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9359440/v1/dc1b99dbb428eb9961868b4f.xlsx"}],"financialInterests":"Competing interest reported. GB declares an advisory relationship with Otsuka Canada, and received honoraria for presentations and unrestricted research funding from Biocon Biologics Canada Inc. SR declares an advisory relationship with Amgen Canada, Kyowa Kirin, and Apotex. SR received speaker fees and funding from Kyowa Kirin, as well as funding from Insmed. HAC declares an advisory relationship with Abbvie, Amgen Canada, AstraZeneca, Celltrion, Eli Lilly \u0026 Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Novartis Canada, Pfizer Canada, Sandoz Canada and Sobi. HAC received speaker fees from Abbvie, Amgen Canada, AstraZeneca, Bristol Myers Squibb, Celltrion, Eli Lilly \u0026 Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Mantra Pharma, Novartis Pharmaceuticals Canada, Pfizer Canada and Sobi. HAC was awarded funding from Abbvie, AstraZeneca, Bristol Myers Squibb, Daiichi Sankyo, Eli Lilly \u0026 Co, Fresenius Kabi USA, Neomed, Novartis Pharmaceuticals Canada, Pfizer Canada, Sanofi and Viela Bio. All other authors declare no known competing interests related to the work reported in this article.","formattedTitle":"Serum mass spectrometry unveil the heterogeneity of rheumatoid arthritis and reveals insights into Complement Pathways and IGHV as potential prognostic markers","fulltext":[{"header":"Introduction:","content":"\u003cp\u003eRheumatoid arthritis (RA) is the most common type of autoimmune inflammatory arthritis, affecting approximately 1% of the global population. Current classification of RA still relies heavily on clinical manifestations combined with serological analysis of autoantibodies, primarily rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPA), and erosiveness. Although the heterogeneity of RA pathogenesis between RF/ACPAs positive and negative serotypes has been highlighted at the level of genetic predisposition\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e, immune cell profile\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e and disease aggressiveness including risk of developing bone erosion\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e, these differences are too subtle to readily inform clinical decision. Furthermore, they are inadequately addressed by the presently available clinical tools. Current management strategies for RA thus primarily aim to provide early and aggressive treatment\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e in order to quickly control inflammation and prevent irreversible joint damage that could lead to disability.\u003c/p\u003e \u003cp\u003eMany disease-modifying antirheumatic drugs (DMARD) are approved for treating RA, with methotrexate being the gold standard and often considered as the first-line therapy for DMARD-naive patients\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Indeed, RA management guidelines by the American College of Rheumatology (ACR) recommend a treat-to-target strategy\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e, which relies on clinicians\u0026rsquo; ability to regularly assess disease activity, monitor treatment response and adjust treatment until remission is reached. In order to make such decisions and to monitor disease activity, several tools are available such as the CDAI, SDAI and DAS28-CRP which are all recommended by the ACR\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. However, disease activity still does not accurately predict erosion risk\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e, especially in patients with low disease activity, where approximately 15% still develop erosion\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e, likely due to underlying subclinical inflammation\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Furthermore, these disease activity measures fail to reliably predict response to treatment\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. To circumvent these imperfections in the assessment of RA patients, a matrix estimating erosion progression risk (AUC 0.68) was developed using four baseline characteristics (RF positivity, X-ray erosions, serum C-reactive protein [CRP] levels, and Swollen Joint Count 28); however this tool still lacks sufficient predictive power (PPV 0.30) for integration into clinical practice\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Therefore, identifying biomarkers able to predict treatment response, erosiveness, or joint damage would allow for more accurate risk depiction and better support clinical decisions in managing RA.\u003c/p\u003e \u003cp\u003eThe multi-biomarker disease activity (MBDA) score is a 12-protein biomarker test that correlates with clinical disease assessment using the DAS28-CRP\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. While CDAI, SDAI and DAS28-CRP scores rely on assessment of RA signs and symptoms, the MBDA offers a more objective evaluation of disease activity, free from the biases of physical examinations and inter-user variability. MBDA scoring has been suggested to correlate with radiographic progression with improved performance when adjusted for age, sex and leptin (a peptide hormone regulating energy balance and body weight) levels\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan additionalcitationids=\"CR19\" citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Nonetheless, baseline MBDA scores have not been shown to reliably predict treatment responses, remission or radiographic progression\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e, leaving its cost-effectiveness and clinical utility unsettled. Interestingly, a recent ambitious plasma proteomic study, in 41,931 participants from the UK Biobank Pharma Proteomics Project, demonstrated that combining proteomic signatures, comprising 5 to 20 proteins, with clinical features could significantly improve risk prediction for various diseases\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. Specifically, protein signatures improved risk prediction in rheumatic diseases such as Gout (ΔC-index: 0.0307, Integrated Discrimination Improvement (IDI): 1.3%), Sj\u0026ouml;gren\u0026rsquo;s (ΔC-index: 0.1178, IDI: 11.78%) and Osteoarthritis (ΔC-index: 0.0235, IDI: 2.5%). This finding supports the potential of proteomic investigations endeavours towards the discovery of protein signatures predictive of RA outcomes.\u003c/p\u003e \u003cp\u003eSeveral proteomic studies have previously investigated biomarkers associated with RA diagnosis\u003csup\u003e\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e or in the pre-clinical RA phase, i.e. in at-risk and pre-progression individuals\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. Previous work from our group has shown that adding high serum levels of 14-3-3η to autoantibodies (ACPAs) and CRP slightly improved prediction of impending rapid erosion progression in our longitudinal Early Undifferentiated PolyArthritis (EUPA) cohort\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. Moreover, our group provided proof-of-concept evidence that miRNA profiles derived from PBMC-generated osteoclasts can enhance the identification of rheumatoid arthritis (RA) patients at risk of developing erosions, with predictive models achieving an accuracy of 78% and an area under the curve (AUC) of 0.85; however, this approach is not yet clinically available\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. In the current study, we leveraged the superior dynamic range and deep coverage of Data-Independent Acquisition (DIA) mass spectrometry\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e to overcome the analytical challenges associated with blood proteomic analysis\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e, enabling us to comprehensively profile the serum protein landscape in untreated and treated RA patients. We took an exploratory approach to identify serum protein biomarkers with potential to predict RA disease activity and erosion outcomes.\u003c/p\u003e"},{"header":"Material and methods:","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePatient Cohort:\u003c/h2\u003e \u003cp\u003ePatients were selected from the longitudinal Early Undifferentiated PolyArthritis (EUPA; NCT00512239) cohort at the CIUSSS de l\u0026rsquo;Estrie-CHUS, Canada. This study used serum samples collected at the baseline and the 12-month follow-up visits from previously DMARD-na\u0026iuml;ve RA patients enrolled in the EUPA cohort described elsewhere\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e,\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Patients were selected based on their disease activity status at 12 months, with DAS28-CRP scores\u0026thinsp;\u0026ge;\u0026thinsp;3.2 indicating active disease and \u0026le;\u0026thinsp;2.6 indicating remission\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Additionally, we considered the erosion outcome at the 30- to 60-month visit, categorizing patients as erosive (Sharp erosion score\u0026thinsp;\u0026ge;\u0026thinsp;5) or non-erosive (Sharp erosion score\u0026thinsp;\u0026lt;\u0026thinsp;5)\u003csup\u003e34\u003c/sup\u003e. Patients were extracted from our biobank in order to equilibrate both seropositivity and persistence of activity at the 12-month follow up visit between the two groups. The initial patient subset consisted of CCP+/RF+ individuals, divided into four groups of 10 (n\u0026thinsp;=\u0026thinsp;40): active/erosive, active/non-erosive, remission/erosive, and remission/non-erosive (Supp. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). A subsequent batch included 22 patients who were either CCP\u0026thinsp;+\u0026thinsp;or RF+, with 12 in an active state and 10 in remission; the remaining 50 participants were seronegative (CCP-/RF-), comprising 25 active and 25 in remission. Patient selection was extended to seronegative patients to broaden the scope of our analysis to identify biomarkers of severity independent of serology status. Indeed, our group has previously shown that serological status has lost most of its predictive value in modern cohorts of RA\u003csup\u003e33\u003c/sup\u003e, hence the need for more reliable biomarkers and the expended analysis to seronegative patient. In addition, this extension allowed us to explore potential differences in serum proteome specific to seropositive or seronegative patients. Serum samples were processed following standard operating procedures of the EUPA cohort. Briefly, blood collected from individuals was kept at 4\u0026deg;C (up to a maximum of 5 days) until samples were processed; serum was then prepared, aliquoted, and frozen at -20\u0026deg;C\u003csup\u003e35\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMass Spectrometry (MS):\u003c/h3\u003e\n\u003cp\u003eSample preparation:\u003c/p\u003e \u003cp\u003ePrior to serum processing for mass spectrometry analysis, the protein content was quantified using BioRad Protein Assay Dye (#5000006) and 50 \u0026micro;g of patients' serum protein were resuspended in Urea buffer (8 M urea, 10 mM Hepes, 1 M NH4HCO3) to a final volume of 50 \u0026micro;l. Samples were processed as follows: proteins were first reduced by adding DTT to a final concentration of 5 mM, followed by heating for 2 min at 95\u0026deg;C and an additional incubation at room temperature (RT) for 30 min. The reduced protein samples were then alkylated by adding chloroacetamide to a final concentration of 7.5 mM and incubating for 20 min in the dark at RT. Subsequently, proteins were digested overnight with 1 \u0026micro;g of trypsin (Promega; V5280) at 30\u0026deg;C. The following day, trypsin digestion was stopped by addition of trifluoroacetic acid (TFA) to a final concentration of 0.2% and the generated peptides were desalted and cleaned using ZipTip C18 columns (Millipore Sigma). Peptide desalting was performed as follows: ZipTips pre-wetted in 100% acetronitrile (ACN) were equilibrated with 0.1% TFA, then samples were loaded by repeated up and down pipetting, washed with 0.1% TFA, and eluted in 1% formic acid (FA)/50% ACN. Desalted samples were then vacuum dried and resuspended in 1% FA.\u003c/p\u003e \u003cp\u003eData acquisition:\u003c/p\u003e \u003cp\u003eFor each sample, 250 ng of peptides were injected in a HPLC (nanoElute, Bruker) and loaded onto a trap column at a flow of 4 \u0026micro;l/min (Acclaim PepMap100 C18 column, 0.3 mm id \u0026times; 5 mm, Dionex Corp.) then eluted onto an analytical C18 column (1.9 \u0026micro;m beads, 75 \u0026micro;m \u0026times; 25 cm, PepSep). Peptides were eluted over a 2-h gradient of acetonitrile (5\u0026ndash;37%) in 0.1% FA at 400 nl/min, while being injected into a TimsTOF Pro ion mobility mass spectrometer equipped with a Captive Spray nano electrospray source (Bruker). Data were acquired as follows: for each single TIMS (100 ms) in data-independent acquisition and parallel accumulation serial fragmentation (diaPASEF) mode, we used one mobility window consisting of 27 mass steps (m/z between 114 to 1414 with a mass width of 50 Da) per cycle (1.27 sec. duty cycle). These steps cover the diagonal scan line for +\u0026thinsp;2 and +\u0026thinsp;3 charged peptides in the m/z-ion mobility plane.\u003c/p\u003e \u003cp\u003eProtein identification and quantification:\u003c/p\u003e \u003cp\u003eProtein identifications were performed on raw TimTOF data using DIA-NN v1.8.1 software\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e and the Uniprot human reference proteome (UP000005640_9606). An \u003cem\u003ein silico\u003c/em\u003e spectral library was generated from peptides from our DIA runs. Default DIA-NN settings were used with the following parameters: 1 missed cleavage allowed, cysteine carbamidomethylation was set as fixed modification, Mass and MS1 accuracy were set at 20 ppm, and Min/Max precursor charge were set at 2 and 4. The MaxLFQ-based quantification was performed with a precursor q value filtering at 0.05\u003csup\u003e37\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eData processing:\u003c/p\u003e \u003cp\u003eWe used ProStaR/DAPAR (Wieczorek S. et al. 2017) software\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e and loaded the Label-Free Quantitation (LFQ) protein intensities obtained from DIA-NN. The LFQ intensities were log2-transformed and filtered to retain only proteins identified by 2 or more unique peptides. Only the proteins that had less than 30% of Partially Observed Values (POV) in both baseline and 12 months post-treatment condition were kept. Then, normalization was performed by quantile centering (15% quantile) and missing values were imputed using DetQuantile algorithm for both POV and Missing in the Entire Condition (MEC) with Quantile\u0026thinsp;=\u0026thinsp;5% and Factor\u0026thinsp;=\u0026thinsp;0.2.\u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis:\u003c/h2\u003e \u003cp\u003eR software v4.3.2\u003csup\u003e39\u003c/sup\u003e was used to perform data processing, statistical analyses and visualization, with the following packages: cluster, cowplot, dplyr, ggforce, ggfortify, ggplot2, ggrepel, ggsankey, gt, purrr, stringr, tibble, and tidyr.\u003c/p\u003e \u003cp\u003ePatient clustering analysis.\u003c/p\u003e \u003cp\u003eWe performed two separate clustering analyses to assess unbiased RA patients\u0026rsquo; aggregation based on their serum proteome profiles at baseline and 12-month follow-up visit. The optimal number of clusters was evaluated using Silhouette width and a consensus of multiple methods. We proceeded using the Hierarchical Clustering on Principal Components (HCPC) method of the FactoMineR and factoextra R packages\u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Briefly, Principal Component Analysis (PCA) analyses were conducted on both baseline and post-treatment serum proteome data, followed by HCPC clustering using Euclidean distance metrics and Ward\u0026rsquo;s method, with a maximum of 100 iterations for consolidation. Differently abundant proteins (DAP) between patient clusters were identified by Mann-Whitney U test with Bonferroni correction. The clinical and demographic patient data comparisons between clusters were conducted using the Mann-Whitney U test for continuous values or either the Chi-square test of independence or Fisher\u0026rsquo;s exact test for categorical values. Protein levels were normalized using z-scores calculated for each protein as (value \u0026ndash; mean of all patients)/standard deviation of all patients, for baseline and 12-month proteomes. This allowed comparison of multiple protein serum levels across the patient cohort.\u003c/p\u003e \u003cp\u003eLongitudinal serum proteome analysis.\u003c/p\u003e \u003cp\u003eFor repeated measure analysis, Generalized Estimating Equations (GEE) models were performed using the geepack R package\u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. GEE models were used to evaluate association of serum proteins levels (baseline and post-treatment) with binomial outcomes of disease activity (DA) or erosion status at the 12- to 30-month visits as the response variables. A GEE model was performed for each protein as a predictor term with the following additional predictors as potential confounders: age, gender, serology and symptom duration. The blood draw timeline (baseline vs post-treatment) was used as the repeated measure (wave) with a binomial family of distribution. GEE models performed for DA outcomes used an \u0026ldquo;exchangeable\u0026rdquo; (compound symmetry) correlation structure to leverage DA outcomes measured both at baseline and post-treatment at the 12 months visit, while an \u0026ldquo;independent\u0026rdquo; structure was used for analysis of erosion outcome. GEE analyses were repeated using serology (ACPA/RF)-specific patient subgroups or all patients combined. However, a limited number of patients were excluded from GEE models due to the following missing values: symptom duration NA (1), Sharp score NA (1), baseline DAS28-CRP NA (4), and baseline DAS28-CRP between 2.6\u0026ndash;3.2 (3). The Odds Ratios (OR) with a 95% Confidence Interval (CI) and p values were extracted from the GEE models using easystats package\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e and visualized in volcano plots.\u003c/p\u003e \u003cp\u003eBaseline serum proteome analysis.\u003c/p\u003e \u003cp\u003eLogistic regression models were conducted to identify potential biomarkers predictive of DA or erosion outcomes using the generalized linear models (glm) function with a binomial (\"logit\") family of distribution from the \u0026ldquo;stats\u0026rdquo; R package. Models were performed with baseline serum protein levels as predictor terms, also including age, gender, serology, and symptom duration confounders as additional terms for DA or erosion binomial outcomes as response terms. The Odds Ratios (OR) with a 95% Confidence Interval (CI) and p values were extracted from the models using easystats package, then visualized in Forest plots. Subsequently, the selection of predictive features was refined by filtering proteins which had fold changes (FC) above 1.25X or below 0.75X in the baseline serum proteome.\u003c/p\u003e \u003cp\u003eModel performance assessment\u003c/p\u003e \u003cp\u003eReceiver Operating Characteristic (ROC) curve analyses were performed on logistic regression-selected proteins to evaluate their ability to discriminate between DA or erosion outcomes. ROC curves were analyzed using the predicted probabilities from univariate and multivariate logistic regression models using the pROC R package\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. For multivariate logistic regression models, combinations of predictor terms up to 4 for DA or 3 for erosion (due to lower n) from selected proteins were tested, with additional terms for age, serology, and symptom duration confounders. Predictor combinations were filtered using ROC curve AUC values to retain only those combinations where the addition of a predictor term increased AUC values by a minimum threshold of 0.01. To further refine our results, models were filtered using a penalized-AUC value based on the number of predictors to select the most optimal predictor combinations. The penalized-AUC value was calculated using a threshold of 0.04 per additional predictor terms. Models were ranked based on their penalized AUC and positive predictive values (PPV) values, which were weighted equally (1:1), and the top-performing models were selected. DeLong\u0026rsquo;s test was used to compare the performance of the identified predictors, allowing to determine if they contributed statistically significant improvements.\u003c/p\u003e \u003cp\u003eKEGG pathway enrichment analysis.\u003c/p\u003e \u003cp\u003eKEGG pathway enrichment analyses were carried out with active subnetworks using the PathfindR R package\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e,\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e, which leverage Protein-protein Interaction Network (PIN) information. KEGG enrichment analyses were performed using proteins and p values from our GEE results as inputs, the KEGG gene set, the Biogrid PIN, the Benjamini \u0026amp; Hochberg p adjustment and enrichment threshold of 0.05 which was repeated over 10 iterations. Similarly, KEGG pathway analysis was also conducted on DAPs between patient clusters using Bonferroni adjusted p values from Mann-Whitney U test and log2 fold changes.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results:","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\n \u003ch2\u003eCohort description and experimental design\u003c/h2\u003e\n \u003cp\u003eSerum proteomic profiling was performed on a selection of 48 seronegative and 59 seropositive RA patients previously enrolled in the EUPA longitudinal cohort\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e between 2005 and 2019, described in Table\u0026nbsp;1. The selected patients were all DMARD-na\u0026iuml;ve at inclusion and consisted of individuals with disease activity (DA) levels at the 12-month visit assessed as active or in remission in a balanced ratio. The initial selection of 40 seropositive patients incorporated an additional criterion of erosive status at the 30-60-month visits, however, this criterion was removed for subsequent patient selections to expand the cohort (Supp. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The demographics, including age, sex, body mass index (BMI), were similar among our seronegative and seropositive patient subgroups. RA patients from both groups showed comparable baseline disease activity (SJC68, TJC66, DAS28-CRP, SDAI, CDAI). By the 12-month time point, seropositive and seronegative patients had been exposed to DMARDs for a median of 6.8 months (IQR; 4.9\u0026ndash;8.5) and 7 months (IQR; 5.2\u0026ndash;8.8), respectively. DIA mass spectrometry serum profiling enabled identification of 869 proteins from which 368 were retained following quality control and filtering (see Methods) for downstream statistical analyses.\u003c/p\u003e\n \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eExploring RA patients clustering based on serum proteome landscapes.\u003c/span\u003e\u003c/p\u003e\n \u003cp\u003eWe began by evaluating whether serum proteome profiles could identify distinct RA patient subgroups, using clustering analyses at baseline and post-treatment (12-month visit). There were no confounding effects of therapy at baseline, while the 12-month visit accounted for longitudinal changes of protein expression patterns under treatment. Hopkins\u0026rsquo; statistic values of 0.666 when using baseline serum proteome and of 0.665 for the post-treatment serum indicated only moderate clustering tendencies. We thus applied Hierarchical Clustering on Principal Components (HCPC) to separate patients in two clusters based on serum proteomic profiles at baseline (Fig. \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA) and post-treatment (Fig. \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). This clustering reflected the optimal number of clusters suggested by the silhouette width (Supp. Figure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA,C), as well as by the majority of clustering metrics (Supp. Figure \u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB,D).\u003c/p\u003e\n \u003cp\u003eComparison of clinical and demographic features across clusters from both baseline and post-treatment proteome clustering did not reveal significant associations with clinical manifestations or outcomes (Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Nonetheless, cluster B1 exhibited a higher proportion of patients seropositive for either RF or ACPAs (69%) compared to cluster B2 (43%, p\u0026thinsp;=\u0026thinsp;0.011) (Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e; Supp. Figure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Disease activity measures at the 12-month visit, such as the DAS28-CRP, SDAI and HAQ, were comparable across clusters at baseline (Supp. Figure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC) and post-treatment (Supp. Figure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). Furthermore, patients across clusters had similar CRP blood levels, a surrogate of inflammation, both at baseline or at the 12-month visit (Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e). There was also no clear association of DA (Supp. Figure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE-F) or erosion (Supp. Figure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG-H) outcomes with patients clustering.\u003c/p\u003e\n \u003cp\u003eTo further investigate our clustering results, we analyzed the differently abundant proteins across patient clusters by performing KEGG enrichment analysis. Notably, the complement and coagulation cascades were the most significantly enriched pathways both in baseline (Supp. Figure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA) and post-treatment (Supp. Figure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB) clustering. Patients\u0026rsquo; serum levels of proteins related to the complement and coagulation cascades pathway, as shown by the mean z-scores, segregated both at baseline (Fig. \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC; Supp. Figure \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e) and post-treatment (Fig. \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD; Supp. Figure 6). Next, we compared patients\u0026rsquo; trajectories between baseline serum-based clusters and corresponding post-treatment serum clusters (Fig. \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). Patients in cluster B1 were split 43% in P1 and 57% in P2, while most patients (88%) of cluster B2 were assigned to P2. Notably, we found significant differences in mean z-scores for complement-related proteins between baseline and 12-month serum samples of patients who transitioned from B1 to P2 (p adj. = 2.8e-5) and B2 to P1 (p adj. = 0.016) clusters (Fig. \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF).\u003c/p\u003e\n \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eIdentification of disease activity associated proteins by longitudinal analysis of RA patient serum proteomes.\u003c/span\u003e\u003c/p\u003e\n \u003cp\u003eFor a longitudinal assessment of proteins correlation with outcomes, we performed Generalized Estimating Equation (GEE) models incorporating both baseline and 12-month measurements. By analyzing both time points simultaneously, we captured the dynamic changes in protein expression that may be associated with disease activity or erosion outcomes. Studies on prognosis of RA patients from the Canadian Early Arthritis Cohort (CATCH) suggested seropositive RA patients showed lower treatment responses and higher risk of erosion\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Thus, we first implemented GEE models separately on seropositive and seronegative patient subsets for each protein with binomial outcomes, then GEE analyses were repeated on complete cohort data.\u003c/p\u003e\n \u003cp\u003eFirst, GEE analyses using serum proteome with DAS28-CRP based binomial outcome identified 39 and 75 disease activity-associated proteins (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in seronegative (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA) and seropositive (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB) subgroups, respectively. Furthermore, analysing the complete RA cohort\u0026rsquo;s serum proteomes regardless of serology revealed 64 proteins significantly correlated with DA (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC). A subset of these proteins exhibited FCs above 1.25X or below 0.75X at either baseline or 12-month, thus most likely to have significant impact, including ADIPOQ, CRP, FAH, IGHV3-43, IGKV2-29; IGKV2D-29, IGLV1-36, IGLV3-27, IGLV8-61, SAA1, SAA2, TAGLN2 and XIRP2. Moreover, there was partial overlap between proteins identified from analyses performed using serology-specific or the full cohort data, with 14 proteins consistently identified across all three GEE analyses (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD). Subsequently, we assessed biological function by conducting KEGG pathway enrichment analysis using DA-associated proteins identified through our GEE analysis of serum proteomes from the full cohort. The top 10 enriched KEGG terms related to RA disease activity were primarily components of the complement system, with the addition of lipoproteins and platelet activation (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE).\u003c/p\u003e\n \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eIdentification of erosion associated proteins by longitudinal analysis of RA patient serum proteomes.\u003c/span\u003e\u003c/p\u003e\n \u003cp\u003eWe applied the same GEE approach to analyze serum proteins associated with erosion status binomial outcomes\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. These GEE analyses of RA patients\u0026rsquo; serum proteomes identified 22 and 44 proteins associated with erosion (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in seronegative (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA) and seropositive (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB) subgroups, respectively. Additionally, 27 proteins were identified as significantly correlated with erosion (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) by analysing the complete RA cohort\u0026rsquo;s serum proteomic data regardless of serology (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). Among these, CA1, CETP, CTBS, DNAH11, ENPP2, IGHV1-18, IGHV1-2, IGHV1-46, IGHV2-26, IGHV5-10-1, IGHV5-51, IGLV3-9, MPO exhibited FCs\u0026thinsp;\u0026ge;\u0026thinsp;1.25X or \u0026le;\u0026thinsp;0.75X at either baseline or 12-month. Overall, there was more limited overlap between erosion-associated proteins identified across GEE analyses using serology-specific subgroups or the full cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD), and none was common in the 3 GEE analyses. Nonetheless, the analysis of KEGG pathways using erosion-associated proteins identified from the complete cohort suggested enrichment of phagosome and neutrophil extracellular trap formation pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE).\u003c/p\u003e\n \u003cp\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eExploring baseline serum profiles to predict RA outcomes.\u003c/span\u003e\u003c/p\u003e\n \u003cp\u003eTo explore biomarkers that may predict rheumatoid arthritis outcomes, including DA at 12-month and erosiveness, we analyzed baseline serum proteomic profiles using generalized linear modeling (GLM) with binomial outcome (logistic regression) approaches, adjusted for age, gender and serology as confounders. For DA outcome at 12-month, we identified eleven proteins with baseline serum levels significantly associated (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Within these proteins, APOC4, PFN1, and SAA1;SAA2 had FCs\u0026thinsp;\u0026ge;\u0026thinsp;1.25X or \u0026le;\u0026thinsp;0.75X (in blue in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). These findings indicate that baseline serum concentrations of these proteins may serve as predictors of resistance to first-line therapies.\u003c/p\u003e\n \u003cp\u003eNext, we analyzed the association between baseline serum protein levels and binomial erosion outcome using GLM models adjusted for age, gender and serology. We observed 17 proteins with baseline serum levels significantly (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) associated with erosion (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). The following 7 proteins had FCs\u0026thinsp;\u0026ge;\u0026thinsp;1.25X or \u0026le;\u0026thinsp;0.75X : CA1, CTBS, IGHV1-18, IGHV1-46, IGHV2-26, IGHV5-10-1;IGHV5-51, and MPO (in blue Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003eEvaluating our protein biomarker candidates\u0026rsquo; performance\u003c/h2\u003e\n \u003cp\u003eTo further evaluate the performance of the identified protein biomarker candidates, we conducted ROC curve analyses to assess their discriminatory power in distinguishing between patients based on 12-month DA or erosion outcomes. For DA, each protein identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA) performed non-inferiorly compared to CRP, a known predictor (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, see also Supp. Tables\u0026nbsp;1 and \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e); SAA1;SAA2 (AUC\u0026thinsp;=\u0026thinsp;0.636, PPV\u0026thinsp;=\u0026thinsp;59.6%) and PFN1 (AUC\u0026thinsp;=\u0026thinsp;0.623, PPV\u0026thinsp;=\u0026thinsp;60%) had slightly higher AUC values compared to CRP (AUC\u0026thinsp;=\u0026thinsp;0.601, PPV\u0026thinsp;=\u0026thinsp;59.1%), a non statistically significant difference. Next, we explored whether combining multiple protein predictors (up to four only to avoid overfitting) could enhance overall performance. The ROC curves from our top five models selected based on balanced improvements of AUC and PPV are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB and their performance were compared using DeLong\u0026rsquo;s test (Supp. Tables\u0026nbsp;1 and \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The predictive performance was further improved when combining proteins, such as SAA1;SAA2 with PFN1 (AUC\u0026thinsp;=\u0026thinsp;0.677, PPV\u0026thinsp;=\u0026thinsp;71.1%) or adding APOC4 to this combination (AUC\u0026thinsp;=\u0026thinsp;0.691, PPV\u0026thinsp;=\u0026thinsp;65.9%). Neither the SAA1;SAA2-PFN1 (p\u0026thinsp;=\u0026thinsp;0.13) nor the SAA1;SAA2-PFN1-APOC4 (p\u0026thinsp;=\u0026thinsp;0.09) signatures showed significant superiority over CRP alone (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB, Supp. Tables\u0026nbsp;1 and \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n \u003cp\u003eNext, we assessed the performance of our protein candidates identified in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD as predictors of erosiveness. Notably, autoantibody serology, a known predictor of erosion\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e, exhibited a ROC AUC value of 0.722 (PPV\u0026thinsp;=\u0026thinsp;50%) when adjusted for age and symptom duration as confounders (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC, Supp. Table\u0026nbsp;1). Interestingly, all protein predictors of erosion tested yielded AUC values higher than serology (Supp. Table\u0026nbsp;1), but only IGHV1-18 (AUC\u0026thinsp;=\u0026thinsp;0.832, PPV\u0026thinsp;=\u0026thinsp;61.9%, p\u0026thinsp;=\u0026thinsp;0.009) was statistically superior (Supp. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The ROC curves from the adjusted univariate GLM models from the top 5 proteins with FCs\u0026thinsp;\u0026ge;\u0026thinsp;1.25X or \u0026le;\u0026thinsp;0.75X, along with serology (RF or ACPAs), are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC. Three of these five proteins are related to variable domains of immunoglobulin heavy chains. Given our limited number of erosive patients, we combined only up to 3 protein predictors. The results from predictor signatures were refined by selecting the top 5 protein combinations with the best performance based on both their AUC and PPV values (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). Combining CTBS with IGHV1-18 (AUC\u0026thinsp;=\u0026thinsp;0.846, PPV\u0026thinsp;=\u0026thinsp;68.2%) significantly outperformed serology (p\u0026thinsp;=\u0026thinsp;0.004). Although the CTBS-IGHV1-18 signature had slightly higher PPV compared to IGHV1-18 alone (68.2% vs 61.9%), their AUCs were not statistically different (0.846 vs 0.832, p\u0026thinsp;=\u0026thinsp;0.255). While improved performance over serology was observed for several protein signatures (Supp. Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e2\u003c/span\u003e), these failed short of exhibiting improvements over IGHV1-18 alone.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"Discussion:","content":"\u003cp\u003ePredicting which very early RA patients will respond to specific therapies or will progress to joint erosiveness remain some of the major challenges in contemporary rheumatology, yet are essential for truly personalized RA management. While we were able to capture RA patient heterogeneity through serum proteomes both at baseline and at the 12-month follow-up, clustering patients did not correlate well with clinical outcomes (DAS28-CRP or erosion). This highlights the observable heterogeneity in RA, but also underscores the current limitations in translating proteomic diversity into actionable clinical insights. RA patients in stable remission from the RETRO cohort could be segregated based on proteomic profiles into 4 clusters\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. Similar to our observations, their patient clustering did not clearly associate with future RA outcomes but rather with clinical features. Our current methodology did not allow to assess the impact of these clusters on response to specific treatments. Across studies, the number and the degree of separation of patient clusters using plasma/serum proteomes have greatly varied. Several studies even noted an overlap between clusters of RA patients with clusters composed of healthy controls\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e, while others were able to observe sharper cluster separation\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, such as in Tocilizumab responders and non-responders\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. A more comprehensive analysis, including both distal (serum/plasma) and local (synovial) samples, alongside multi-omic approaches may help refine our understanding of RA patient phenotypes, improve clustering and help establish a link to relevant clinical outcomes.\u003c/p\u003e \u003cp\u003eIn this study, differently abundant serum proteins between patient clusters were enriched for proteins of the complement pathways (Supp. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), despite the absence of association with serology or DA outcomes (Supp. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). CRP blood levels also did not distribute across patient clusters defined by baseline (Supp. Figure\u0026nbsp;7A) or 12-month (Supp. Figure\u0026nbsp;7B) proteomic profiles. This suggests that heterogeneity of circulating complement components is not merely a reflection of inflammation levels but reflects a different polarisation of the immune system in some RA patients. While we did not observe a clear link between serum complement proteins and serology, recent single-cell RNA sequencing data revealed complement cascade gene enrichment in macrophages of ACPA negative compared to ACPA positive patients\u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e. While further work is needed to elucidate a potential serology-specific contribution of the complement pathways, its role in the pathogenesis of RA was previously documented. Previous studies revealed complement pathway terms were enriched in RA patients compared to healthy individuals\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e,\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, associated with DAS28 in stable RA patients\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e, and associated with Tocilizumab response\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. Furthermore, several complement deficiency mouse models demonstrated contribution of the complement pathways to collagen (CIA) and collagen antibody (CAIA) induced arthritis\u003csup\u003e\u003cspan additionalcitationids=\"CR54 CR55 CR56\" citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e. In addition, rheumatoid factor is known to activate complement\u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e,\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e, while C5a, an anaphylatoxin, seems to be the primary complement activation product driving tissue damage in RA, though membrane attack complex deposition\u003csup\u003e\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e\u003c/sup\u003e and C3b-mediated opsonization\u003csup\u003e\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e also play significant roles. A more comprehensive characterization of complement proteins involved in RA by integrating analysis of the synovium, independently of serological profiles, may offer valuable insights for identifying patients who could benefit from complement system modulators such as avacopan, a C5aR1 antagonist\u003csup\u003e\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u003c/sup\u003e, highlighting complement modulation as an emerging therapeutic option under investigation for rheumatic diseases\u003csup\u003e\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eOur cohort of RA patients had similar disease activity at baseline, decreasing variability in protein expression and allowing the identification of subtler changes. Consequently, we did not apply false discovery rate (FDR) correction in GEE and GLM analyses using binomial outcomes, enabling identification of a broader set of proteins more suitable for network enrichment analysis. While this exploratory approach increases the risk of false-positives, 29 proteins remained significantly associated with DA outcome after applying Storey's q-value FDR correction (Supp. Figure\u0026nbsp;8A)\u003csup\u003e\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e. These proteins were also consistently enriched for similar KEGG pathways (Supp. Figure\u0026nbsp;8B). In contrast, only IGHV1-18 was retained as a significant predictor of erosion outcome following FDR correction, likely due to reduced statistical power from smaller and imbalanced comparison groups of erosive patients. Our longitudinal analysis aligns with previous reports showing elevated serum levels of inflammation markers, CRP, SAA1 and SAA2 in RA patients and their association with disease progression\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan additionalcitationids=\"CR66 CR67\" citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e\u003c/sup\u003e. Perhaps not surprisingly, CRP and SAA, both components of the MBDA score, have been associated with remission, thus underscoring their relevance in RA pathogenesis\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e\u003c/sup\u003e. We also observed association of LRG1 serum levels with disease activity, validating recent reports characterizing LRG1 as a novel inflammation marker correlating with DAS28-CRP/ESR\u003csup\u003e70,71\u003c/sup\u003e. Several proteins associated with DA also correlated with CRP levels, a surrogate marker of inflammation, though with varying strengths and directions (Supp. Figure\u0026nbsp;9). Among the other proteins associated with DA outcome, several are linked to the complement system from the Lectin (FCN3, MASP1) and the classical (C1QB) pathways as well as with the common effector C9. However, circulating levels of complement components by themselves may not fully inform on their activation and pathogenic contribution, particularly within the joint tissues.\u003c/p\u003e \u003cp\u003eAnalyses focused on the baseline serum proteomic profiles of treatment-na\u0026iuml;ve RA patients revealed that APOC4, PFN1, and SAA1;SAA2 exhibited predictive capabilities for the 12-month DA outcomes. Combining these proteins showed a trend toward improved performance over CRP, supporting the potential for serum protein signatures to improve current clinical biomarkers.\u003c/p\u003e \u003cp\u003eWe also identified novel potential predictors of RA erosiveness, including baseline serum levels of MPO and several immunoglobulin variable chain proteins, particularly IGHV1-18. Immunoglobulin chains, including IGHV1-18, have been previously reported as enriched in RA patient exosomes, highlighting their putative role in RA\u003csup\u003e72\u003c/sup\u003e. Recently, the frequency of usage of several V/D/J genes of IGHV was reported to associate with disease activity in RA and SLE\u003csup\u003e\u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e\u003c/sup\u003e. Their relative overabundance might suggest that a group of patients develops an oligoclonal response to one or a small subset of autoantigens. The target antigen(s) of these clones remains unknown but might be a subset of citrullinated proteins, such as citrullinated vimentin, whose corresponding autoantibodies are associated with aggressive course of RA\u003csup\u003e32\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTaken together, these findings suggest that the identification of patients at risk of poor treatment responses or erosive disease through protein signatures may be possible. Moreover, the identification of the targeted antigens might lead to a more precise understanding of RA pathophysiology and to the development of better diagnostic and prognostic tools. The presence of IGHV1-18, or potentially the preferential use of other specific B cells receptor chain, may identify a subset of B cell receptor clones that preferentially recognize autoantigens in patients with erosive rheumatoid arthritis. This observation could suggest a novel framework for understanding seropositivity, one that focuses on a subset of B cells predisposed for autoreactivity, independent of their classical targets such as ACPA or carbamylated proteins.\u003c/p\u003e \u003cp\u003eOur DIA mass spectrometry study design intentionally avoided the removal of highly abundant proteins to maximize the number of patients analyzed longitudinally. Consequently, this study may be limited by its depth of serum proteomic profiling, particularly for low-abondance proteins, due to the challenges with mass spectrometry analyses of high dynamic ranges of plasma and serum protein levels\u003csup\u003e\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e\u003c/sup\u003e. Also, further validation of the identified outcome predictors in an independent cohort is required. Optimised protein signatures have the potential for early tailoring of treatment strategies for RA patients at risk of aggressive disease progression and ultimately improve their quality of life.\u003c/p\u003e"},{"header":"Conclusion:","content":"\u003cp\u003eThis longitudinal serum proteomic profiling of 107 patients led to the identification of novel protein candidates as potential biomarkers of RA outcomes. We analyzed baseline serum proteome from treatment-na\u0026iuml;ve RA patients, thereby avoiding any confounding effects of DMARDs on serum protein levels. While these analyses enabled clustering of patients based on their serum proteomic profiles, we were unable to establish clear links between patient clusters and clinical outcomes. Nonetheless, we identified protein signatures that outperformed classic serologic prediction of RA erosiveness. These findings highlight the complexity and heterogeneity of RA and support the need for further in-depth characterization using multi-omic approaches.\u003c/p\u003e"},{"header":"Abbreviations:","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eACN\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eacetonitrile\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eACPA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eanti-citrullinated protein antibodies\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eACR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eAmerican College of Rheumatology\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eAUC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003earea under the curve\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eBMI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ebody mass index\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCAIA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecollagen antibody\u0026ndash;induced arthritis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCATCH\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eCanadian Early Arthritis Cohort\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCCP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecyclic citrullinated peptide\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCDAI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eclinical disease activity index\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCHUS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecentre hospitalier universitaire de Sherbrooke\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCIA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecollagen-induced arthritis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCIUSSS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ecentre int\u0026eacute;gr\u0026eacute; universitaire de sant\u0026eacute; et de services sociaux\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCRP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eC-reactive protein\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDAP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003edifferentially abundant proteins\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDAPAR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003edifferential analysis of protein abundance with R\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDAS28-CRP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003edisease activity score 28 using CRP\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDIA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003edata-independent acquisition\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ediaPASEF\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003edata-independent parallel accumulation\u0026ndash;serial fragmentation\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDMARD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003edisease-modifying antirheumatic drug\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDTT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003edithiothreitol\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eESR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eerythrocyte sedimentation rate\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEUPA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEarly Undifferentiated PolyArthritis cohort\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFDR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003efalse discovery rate\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGEE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egeneralized estimating equations\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGLM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003egeneralized linear model\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHAQ\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ehealth assessment questionnaire\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHCPC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ehierarchical clustering on principal components\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHPLC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ehigh-performance liquid chromatography\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIGHV\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eimmunoglobulin heavy chain variable region\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIQR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003einterquartile range\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eKEGG\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eKyoto Encyclopedia of Genes and Genomes\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLFQ\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003elabel-free quantification\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003em/z\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emass-to-charge ratio\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMBDA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emulti-biomarker disease activity\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMEC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emissing in entire condition\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003emiRNA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emicroRNA\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePBMC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eperipheral blood mononuclear cells\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePCA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eprincipal component analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePIN\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eprotein interaction network\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePOV\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003epartially observed values\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePPV\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003epositive predictive value\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003erheumatoid arthritis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRF\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003erheumatoid factor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRNA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eribonucleic acid\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eROC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ereceiver operating characteristic\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSAA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eserum amyloid A\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSDAI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003esimplified disease activity index\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSJC68\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eswollen joint count 68\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSLE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003esystemic lupus erythematosus\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTFA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etrifluoroacetic acid\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTIMS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etrapped ion mobility spectrometry\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eTJC66\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003etender joint count 66\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations:","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePatients enrolled in the Early Undifferentiated PolyArthritis (EUPA) cohort gave their informed consent and study was approved by the Ethics Review Board of the CIUSSS de l\u0026rsquo;Estrie-CHUS (ClinicalTrials.gov ID: NCT00512239).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe proteomic datasets generated during this study are not publicly available because patient consent for sharing individual proteomic data was not obtained.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGB declares an advisory relationship with Otsuka Canada, and received honoraria for presentations and unrestricted research funding from Biocon Biologics Canada Inc. SR declares an advisory relationship with Amgen Canada, Kyowa Kirin, and Apotex. SR received speaker fees and funding from Kyowa Kirin, as well as funding from Insmed. HAC declares an advisory relationship with Abbvie, Amgen Canada, AstraZeneca, Celltrion, Eli Lilly \u0026amp; Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Novartis Canada, Pfizer Canada, Sandoz Canada and Sobi. HAC received speaker fees from Abbvie, Amgen Canada, AstraZeneca, Bristol Myers Squibb, Celltrion, Eli Lilly \u0026amp; Co, Fresenius Kabi USA, GSK, Hoffmann-LaRoche, Janssen Pharmaceuticals, Mantra Pharma, Novartis Pharmaceuticals Canada, Pfizer Canada and Sobi. HAC was awarded funding from Abbvie, AstraZeneca, Bristol Myers Squibb, Daiichi Sankyo, Eli Lilly \u0026amp; Co, Fresenius Kabi USA, Neomed, Novartis Pharmaceuticals Canada, Pfizer Canada, Sanofi and Viela Bio. All other authors declare no known competing interests related to the work reported in this article.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe EUPA cohort was supported by the Canadian Institutes for Health Research MOP-110959. We also acknowledge previous support from The Arthritis Society Grants 00/201 and RG06/108. From July 2007 to May 2025, the EUPA cohort also received funding from the Canadian ArThritis CoHort. SR, GB and HAC are part of the Centre de Recherche du Centre Hospitalier Universitaire de Sherbrooke (CR-CHUS), which received a team funding from the Fonds de Recherche du Qu\u0026eacute;bec\u0026ndash;Sant\u0026eacute; (FRQS). FMB, SR, GB and HAC received support from an internal grant from CR-CHUS for Projet structurants en recherche translationnelle. HAC is a Clinical Research Scholar \u0026ndash; Junior 2 from FRQS (https://doi.org/10.69777/369863). HAC is the current chairholder (2\u003csup\u003end\u003c/sup\u003e mandate) of the Chaire Andr\u0026eacute; Lussier de rhumatologie and is supported by the Canadian Institutes of Health Research (452873).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHAC secured the funding for the project. BM and EB processed serum samples for mass spectrometry. DL analyzed the samples on the mass spectrometer and performed peptide/protein identification. DL and FMB contributed to mass spectrometry methodology. NC, SR, JM, GB and HAC were involved in patient enrollment and biobanking. BM and BR performed data analysis. NC contributed to statistical methodology and biobank coordination. BM, NC, BR, SR, JM, GB and HAC revised the manuscript. BM and HAC wrote the original draft and prepared the reviewed manuscript for submission.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe are grateful for the contribution of rheumatologists Drs Guylaine Arsenault, Philippe Bilodeau, Lyne Bissonnette, Alessandra Bruns, Pierre Dagenais, Patrick Liang and Ariel Masetto to EUPA patient recruitment and follow-up. We extend our gratitude to research assistants Chantal Guillet, No\u0026eacute;mie Poirier and Christine Rosa involved in sample biobanking of EUPA patients and at the Banque de Pathologies et Perturbations Immunes et Inflammatoires (BPPII) biobank, located at the Rheumatology Clinic of the CIUSSS de l\u0026rsquo;Estrie-CHUS in Canada. We express our sincere thanks to all the patients who participated in the EUPA study.\u003c/p\u003e"},{"header":"References:","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eViatte S, Plant D, Bowes J, Lunt M, Eyre S, Barton A, Worthington J. Genetic markers of rheumatoid arthritis susceptibility in anti-citrullinated peptide antibody negative patients. Ann Rheum Dis. 2012;71(12):1984\u0026ndash;90. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/annrheumdis-2011-201225\u003c/span\u003e\u003cspan address=\"10.1136/annrheumdis-2011-201225\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu X, Liu Y, Jin S, Wang M, Jiao Y, Yang B, Lu X, Ji X, Fei Y, Yang H, Zhao L, Chen H, Zhang Y, Li H, Lipsky PE, Tsokos GC, Bai F, Zhang X. Single-cell sequencing of immune cells from anticitrullinated peptide antibody positive and negative rheumatoid arthritis. Nat Commun. 2021;12(1):4977. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-021-25246-7\u003c/span\u003e\u003cspan address=\"10.1038/s41467-021-25246-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarra L, Pope JE, Orav JE, Boire G, Haraoui B, Hitchon C, Keystone EC, Thorne JC, Tin D, Bykerk VP, the CATCH Investigators. Prognosis of Seronegative Patients in a Large Prospective Cohort of Patients with Early Inflammatory Arthritis. J Rhuematol. 2014;41(12):2361\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3899/jrheum.140082\u003c/span\u003e\u003cspan address=\"10.3899/jrheum.140082\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeyer O. Anticitrullinated protein/peptide antibody assays in early rheumatoid arthritis for predicting five year radiographic damage. Ann Rheum Dis. 2003;62(2):120\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/ard.62.2.120\u003c/span\u003e\u003cspan address=\"10.1136/ard.62.2.120\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFarragher TM, Lunt M, Fu B, Bunn D, Symmons DPM. Early treatment with, and time receiving, first disease-modifying antirheumatic drug predicts long-term function in patients with inflammatory polyarthritis. Ann Rheum Dis. 2010;69(4):689\u0026ndash;95. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/ard.2009.108639\u003c/span\u003e\u003cspan address=\"10.1136/ard.2009.108639\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSingh JA, Saag KG, Bridges SL, Akl EA, Bannuru RR, Sullivan MC, Vaysbrot E, McNaughton C, Osani M, Shmerling RH, Curtis JR, Furst DE, Parks D, Kavanaugh A, O\u0026rsquo;Dell J, King C, Leong A, Matteson EL, Schousboe JT, McAlindon T. 2015 American College of Rheumatology Guideline for the Treatment of Rheumatoid Arthritis. Arthritis Care Res. 2016;68(1):1\u0026ndash;25. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/acr.22783\u003c/span\u003e\u003cspan address=\"10.1002/acr.22783\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFraenkel L, Bathon JM, England BR, St.Clair EW, Arayssi T, Carandang K, Deane KD, Genovese M, Huston KK, Kerr G, Kremer J, Nakamura MC, Russell LA, Singh JA, Smith BJ, Sparks JA, Venkatachalam S, Weinblatt ME, Al-Gibbawi M, Akl EA. 2021 American College of Rheumatology Guideline for the Treatment of Rheumatoid Arthritis. Arthritis Rheumatol. 2021;73(7):1108\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/art.41752\u003c/span\u003e\u003cspan address=\"10.1002/art.41752\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEngland BR, Tiong BK, Bergman MJ, Curtis JR, Kazi S, Mikuls TR, O\u0026rsquo;Dell JR, Ranganath VK, Limanni A, Suter LG, Michaud K. 2019 Update of the American College of Rheumatology Recommended Rheumatoid Arthritis Disease Activity Measures. Arthritis Care Res. 2019;71(12):1540\u0026ndash;55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/acr.24042\u003c/span\u003e\u003cspan address=\"10.1002/acr.24042\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHambardzumyan K, Bolce RJ, Saevarsdottir S, Forslind K, Wallman JK, Cruickshank SE, Sasso EH, Chernoff D, van Vollenhoven RF. Association of a multibiomarker disease activity score at multiple time-points with radiographic progression in rheumatoid arthritis: Results from the SWEFOT trial. RMD Open. 2016;2(1):e000197. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/rmdopen-2015-000197\u003c/span\u003e\u003cspan address=\"10.1136/rmdopen-2015-000197\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMolenaar ETH, Voskuyl AE, Dinant HJ, Bezemer PD, Boers M, Dijkmans BAC. Progression of radiologic damage in patients with rheumatoid arthritis in clinical remission. Arthr Rhuem. 2004;50(1):36\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/art.11481\u003c/span\u003e\u003cspan address=\"10.1002/art.11481\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCohen G, Gossec L, Dougados M, Cantagrel A, Goupille P, Daures J, Rincheval N, Combe B. Radiological damage in patients with rheumatoid arthritis on sustained remission. Ann Rheum Dis. 2007;66(3):358\u0026ndash;63. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/ard.2006.057497\u003c/span\u003e\u003cspan address=\"10.1136/ard.2006.057497\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGandjbakhch F, Conaghan PG, Ejbjerg B, Haavardsholm EA, Foltz V, Brown AK, M\u0026oslash;ller D\u0026oslash;hn U, Lassere M, Freeston J, B\u0026oslash;yesen P, Bird P, Fautrel B, Hetland ML, Emery P, Bourgeois P, H\u0026oslash;rslev-Petersen K, Kvien TK, McQUEEN F, \u0026Oslash;stergaard M. Synovitis and Osteitis Are Very Frequent in Rheumatoid Arthritis Clinical Remission: Results from an MRI Study of 294 Patients in Clinical Remission or Low Disease Activity State. J Rhuematol. 2011;38(9):2039\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3899/jrheum.110421\u003c/span\u003e\u003cspan address=\"10.3899/jrheum.110421\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCapelusnik D, Aletaha D. Baseline predictors of different types of treatment success in rheumatoid arthritis. Ann Rheum Dis. 2022;81(2):153\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/annrheumdis-2021-220853\u003c/span\u003e\u003cspan address=\"10.1136/annrheumdis-2021-220853\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuedders BA, Johnson TM, Sayles H, Thiele GM, Mikuls TR, O\u0026rsquo;Dell JR, England BR. Predictive ability, validity, and responsiveness of the multi-biomarker disease activity score in patients with rheumatoid arthritis initiating methotrexate. Semin Arthritis Rheum. 2020;50(5):1058\u0026ndash;63. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.semarthrit.2020.05.019\u003c/span\u003e\u003cspan address=\"10.1016/j.semarthrit.2020.05.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVanier A, Smolen JS, Allaart CF, Van Vollenhoven R, Verschueren P, Vastesaeger N, Saevarsdottir S, Visser K, Aletaha D, Combe B, Fautrel B. An updated matrix to predict rapid radiographic progression of early rheumatoid arthritis patients: Pooled analyses from several databases. Rheumatology. 2020;59(8):1842\u0026ndash;52. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/rheumatology/kez542\u003c/span\u003e\u003cspan address=\"10.1093/rheumatology/kez542\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCentola M, Cavet G, Shen Y, Ramanujan S, Knowlton N, Swan KA, Turner M, Sutton C, Smith DR, Haney DJ, Chernoff D, Hesterberg LK, Carulli JP, Taylor PC, Shadick NA, Weinblatt ME, Curtis JR. Development of a Multi-Biomarker Disease Activity Test for Rheumatoid Arthritis. PLoS ONE. 2013;8(4):e60635. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1371/journal.pone.0060635\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0060635\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBakker MF, Cavet G, Jacobs JW, Bijlsma JWJ, Haney DJ, Shen Y, Hesterberg LK, Smith DR, Centola M, van Roon JAG, Lafeber FPJG, Welsing PMJ. Performance of a multi-biomarker score measuring rheumatoid arthritis disease activity in the CAMERA tight control study. Ann Rheum Dis. 2012;71(10):1692\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/annrheumdis-2011-200963\u003c/span\u003e\u003cspan address=\"10.1136/annrheumdis-2011-200963\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCurtis JR, Weinblatt ME, Shadick NA, Brahe CH, \u0026Oslash;stergaard M, Hetland ML, Saevarsdottir S, Horton M, Mabey B, Flake DD, Ben-Shachar R, Sasso EH, Huizinga TW. Validation of the adjusted multi-biomarker disease activity score as a prognostic test for radiographic progression in rheumatoid arthritis: A combined analysis of multiple studies. Arthritis Res Therapy. 2021;23(1):1. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13075-020-02389-4\u003c/span\u003e\u003cspan address=\"10.1186/s13075-020-02389-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCurtis JR, Greenberg JD, Harrold LR, Kremer JM, Palmer JL. Influence of obesity, age, and comorbidities on the multi-biomarker disease activity test in rheumatoid arthritis. Semin Arthritis Rheum. 2018;47(4):472\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.semarthrit.2017.07.010\u003c/span\u003e\u003cspan address=\"10.1016/j.semarthrit.2017.07.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCurtis JR, Flake DD, Weinblatt ME, Shadick NA, \u0026Oslash;stergaard M, Hetland ML, Brahe CH, Hwang YG, Furst DE, Strand V, Etzel CJ, Pappas DA, Wang X, Hwang CC, Sasso EH, Gutin A, Hitraya E, Lanchbury JS. Adjustment of the multi-biomarker disease activity score to account for age, sex and adiposity in patients with rheumatoid arthritis. Rheumatology. 2019;58(5):874\u0026ndash;83. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/rheumatology/key367\u003c/span\u003e\u003cspan address=\"10.1093/rheumatology/key367\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeznerics FA, Kem\u0026eacute;ny LV, Gunther E, Bak\u0026oacute; E, Dembrovszky F, Szab\u0026oacute; B, Ascsill\u0026aacute;n A, Lutz E, Csupor D, Hegyi P, B\u0026aacute;nv\u0026ouml;lgyi A, Nagy G. Multibiomarker disease activity score: An objective tool for monitoring rheumatoid arthritis? A systematic review and meta-analysis. Rheumatology. 2023;62(6):2048\u0026ndash;59. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/rheumatology/keac715\u003c/span\u003e\u003cspan address=\"10.1093/rheumatology/keac715\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarrasco-Zanini J, Pietzner M, Davitte J, Surendran P, Croteau-Chonka DC, Robins C, Torralbo A, Tomlinson C, Gr\u0026uuml;nschl\u0026auml;ger F, Fitzpatrick N, Ytsma C, Kanno T, Gade S, Freitag D, Ziebell F, Haas S, Denaxas S, Betts JC, Wareham NJ, Langenberg C. Proteomic signatures improve risk prediction for common and rare diseases. Nat Med. 2024;30(9):2489\u0026ndash;98. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41591-024-03142-z\u003c/span\u003e\u003cspan address=\"10.1038/s41591-024-03142-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJin L, Wang F, Wang X, Harvey BP, Bi Y, Hu C, Cui B, Darcy AT, Maull JW, Phillips BR, Kim Y, Jenkins GJ, Sornasse TR, Tian Y. Identification of Plasma Biomarkers from Rheumatoid Arthritis Patients Using an Optimized Sequential Window Acquisition of All THeoretical Mass Spectra (SWATH) Proteomics Workflow. Proteomes. 2023;11(4):32. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/proteomes11040032\u003c/span\u003e\u003cspan address=\"10.3390/proteomes11040032\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu C, Dai Z, Xu J, Zhao L, Xu Y, Li M, Yu J, Zhang L, Deng H, Liu L, Zhang M, Huang J, Wu L, Chen G. Proteome Profiling Identifies Serum Biomarkers in Rheumatoid Arthritis. Front Immunol. 2022;13:865425. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2022.865425\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2022.865425\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMun S, Lee J, Park M, Shin J, Lim M-K, Kang H-G. Serum biomarker panel for the diagnosis of rheumatoid arthritis. Arthritis Res Therapy. 2021;23(1):31. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13075-020-02405-7\u003c/span\u003e\u003cspan address=\"10.1186/s13075-020-02405-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO\u0026rsquo;Neil LJ, Spicer V, Smolik I, Meng X, Goel RR, Anaparti V, Wilkins J, El-Gabalawy HS. Association of a Serum Protein Signature With Rheumatoid Arthritis Development. Arthritis Rheumatol. 2021;73(1):78\u0026ndash;88. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/art.41483\u003c/span\u003e\u003cspan address=\"10.1002/art.41483\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO\u0026rsquo;Neil LJ, Meng X, Mcfadyen C, Fritzler MJ, El-Gabalawy HS. Serum proteomic networks associate with pre-clinical rheumatoid arthritis autoantibodies and longitudinal outcomes. Front Immunol. 2022;13:958145. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2022.958145\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2022.958145\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarrier N, De Brum-Fernandes AJ, Liang P, Masetto A, Roux S, Biln NK, Maksymowych WP, Boire G. Impending radiographic erosive progression over the following year in a cohort of consecutive patients with inflammatory polyarthritis: Prediction by serum biomarkers. RMD Open. 2020;6(1):e001191. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/rmdopen-2020-001191\u003c/span\u003e\u003cspan address=\"10.1136/rmdopen-2020-001191\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNguyen HD, Lortie A, Mbous Nguimbus L, Marrugo J, Allard-Chamard H, Bouchard L, Boire G, Scott MS, Roux S. Osteoclast microRNA Profiling in Rheumatoid Arthritis to Capture the Erosive Factor. JBMR Plus. 2023;7(8):e10776. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jbm4.10776\u003c/span\u003e\u003cspan address=\"10.1002/jbm4.10776\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKitata RB, Yang J, Chen Y. Advances in data-independent acquisition mass spectrometry towards comprehensive digital proteome landscape. Mass Spectrom Rev. 2023;42(6):2324\u0026ndash;48. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/mas.21781\u003c/span\u003e\u003cspan address=\"10.1002/mas.21781\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao Y, Xue Q, Wang M, Meng B, Jiang Y, Zhai R, Zhang Y, Dai X, Fang X. Evolution of Mass Spectrometry Instruments and Techniques for Blood Proteomics. J Proteome Res. 2023;22(4):1009\u0026ndash;23. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1021/acs.jproteome.3c00102\u003c/span\u003e\u003cspan address=\"10.1021/acs.jproteome.3c00102\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoire G, Cossette P, de Brum-Fernandes AJ, Liang P, Niyonsenga T, Zhou ZJ, Carrier N, Daniel C, M\u0026eacute;nard H-A. Anti-Sa antibodies and antibodies against cyclic citrullinated peptide are not equivalent as predictors of severe outcomes in patients with recent-onset polyarthritis. Arthritis Res Therapy. 2005;7(3):R592\u0026ndash;603. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/ar1719\u003c/span\u003e\u003cspan address=\"10.1186/ar1719\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarrier N, Roux S, Masetto A, De Brum-Fernandes AJ, Liang P, Maoui M, Boire G. Changes at Presentation in Patients With Early Rheumatoid Arthritis: A 24-Year Study of the Early Undifferentiated Polyarthritis (EUPA) Cohort. J Rhuematol. 2024;52(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3899/jrheum.2024-0560\u003c/span\u003e\u003cspan address=\"10.3899/jrheum.2024-0560\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. jrheum.2024\u0026thinsp;\u0026ndash;\u0026thinsp;0560.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003evan der Heijde D. How to read radiographs according to the Sharp/van der Heijde method. J Rhuematol. 2000;27(1):261\u0026ndash;3.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCarrier N, Marotta A, de Brum-Fernandes AJ, Liang P, Masetto A, M\u0026eacute;nard HA, Maksymowych WP, Boire G. Serum levels of 14-3-3η protein supplement C-reactive protein and rheumatoid arthritis-associated antibodies to predict clinical and radiographic outcomes in a prospective cohort of patients with recent-onset inflammatory polyarthritis. Arthritis Res Therapy. 2016;18(1):37. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13075-016-0935-z\u003c/span\u003e\u003cspan address=\"10.1186/s13075-016-0935-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDemichev V, Messner CB, Vernardis SI, Lilley KS, Ralser M. DIA-NN: Neural networks and interference correction enable deep proteome coverage in high throughput. Nat Methods. 2020;17(1):41\u0026ndash;4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41592-019-0638-x\u003c/span\u003e\u003cspan address=\"10.1038/s41592-019-0638-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCox J, Hein MY, Luber CA, Paron I, Nagaraj N, Mann M. Accurate Proteome-wide Label-free Quantification by Delayed Normalization and Maximal Peptide Ratio Extraction, Termed MaxLFQ. Mol Cell Proteom. 2014;13(9):2513\u0026ndash;26. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1074/mcp.M113.031591\u003c/span\u003e\u003cspan address=\"10.1074/mcp.M113.031591\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWieczorek S, Combes F, Lazar C, Giai Gianetto Q, Gatto L, Dorffer A, Hesse A-M, Cout\u0026eacute; Y, Ferro M, Bruley C, Burger T. DAPAR \u0026amp; ProStaR: Software to perform statistical analyses in quantitative discovery proteomics. Bioinformatics. 2017;33(1):135\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/bioinformatics/btw580\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btw580\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eR Core Team. (2024). \u003cem\u003eR: A Language and Environment for Statistical Computing\u003c/em\u003e [Computer software]. R Foundation for Statistical Computing. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.R-project.org\u003c/span\u003e\u003cspan address=\"https://www.R-project.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL\u0026ecirc; S, Josse J, Husson F. FactoMineR: An \u003cem\u003eR\u003c/em\u003e Package for Multivariate Analysis. J Stat Softw. 2008;25(1). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18637/jss.v025.i01\u003c/span\u003e\u003cspan address=\"10.18637/jss.v025.i01\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKassambara A, Mundt F. (2016). \u003cem\u003efactoextra: Extract and Visualize the Results of Multivariate Data Analyses\u003c/em\u003e (p. 1.0.7) [Dataset]. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.32614/CRAN.package.factoextra\u003c/span\u003e\u003cspan address=\"10.32614/CRAN.package.factoextra\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHalekoh U, H\u0026oslash;jsgaard S, Yan J. The \u003cem\u003eR\u003c/em\u003e Package geepack for Generalized Estimating Equations. J Stat Softw. 2006;15(2). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.18637/jss.v015.i02\u003c/span\u003e\u003cspan address=\"10.18637/jss.v015.i02\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eL\u0026uuml;decke D, Makowski D, Ben-Shachar MS, Patil I, Wiernik BM, Bacher E, Th\u0026eacute;riault R. (2022). \u003cem\u003eeasystats: Framework for Easy Statistical Modeling, Visualization, and Reporting\u003c/em\u003e (p. 0.7.4) [Dataset]. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.32614/CRAN.package.easystats\u003c/span\u003e\u003cspan address=\"10.32614/CRAN.package.easystats\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRobin X, Turck N, Hainard A, Tiberti N, Lisacek F, Sanchez J-C, M\u0026uuml;ller M. pROC: An open-source package for R and S\u0026thinsp;+\u0026thinsp;to analyze and compare ROC curves. BMC Bioinformatics. 2011;12(1):77. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/1471-2105-12-77\u003c/span\u003e\u003cspan address=\"10.1186/1471-2105-12-77\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUlgen E, Ozisik O, Sezerman OU. pathfindR: An R Package for Comprehensive Identification of Enriched Pathways in Omics Data Through Active Subnetworks. Front Genet. 2019;10:858. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fgene.2019.00858\u003c/span\u003e\u003cspan address=\"10.3389/fgene.2019.00858\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanehisa M, Furumichi M, Sato Y, Matsuura Y, Ishiguro-Watanabe M. KEGG: Biological systems database as a model of the real world. Nucleic Acids Res. 2025;53(D1):D672\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/nar/gkae909\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkae909\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBruynesteyn K, Van Der Heijde D, Boers M, Saudan A, Peloso P, Paulus H, Houben H, Griffiths B, Edmonds J, Bresnihan B, Boonen A, Van Der Linden S. Determination of the minimal clinically important difference in rheumatoid arthritis joint damage of the Sharp/van der Heijde and Larsen/Scott scoring methods by clinical experts and comparison with the smallest detectable difference. Arthr Rhuem. 2002;46(4):913\u0026ndash;20. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/art.10190\u003c/span\u003e\u003cspan address=\"10.1002/art.10190\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTakeuchi T, Soen S, Ishiguro N, Yamanaka H, Tanaka S, Kobayashi M, Okubo N, Nitta T, Tanaka Y. Predictors of new bone erosion in rheumatoid arthritis patients receiving conventional synthetic disease-modifying antirheumatic drugs: Analysis of data from the DRIVE and DESIRABLE studies. Mod Rheumatol. 2021;31(1):34\u0026ndash;41. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/14397595.2019.1703484\u003c/span\u003e\u003cspan address=\"10.1080/14397595.2019.1703484\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO\u0026rsquo;Neil LJ, Hu P, Liu Q, Islam MM, Spicer V, Rech J, Hueber A, Anaparti V, Smolik I, El-Gabalawy HS, Schett G, Wilkins JA. Proteomic Approaches to Defining Remission and the Risk of Relapse in Rheumatoid Arthritis. Front Immunol. 2021;12:729681. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2021.729681\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2021.729681\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe Z, Glass MC, Venkatesan P, Feser ML, Lazaro L, Okada LY, Tran NTT, He YD, Zaim R, Bennett S, Ravisankar CE, Dornisch P, Arishi EM, Asamoah NA, Barzideh AG, Becker S, Bemis LA, Buckner EA, Collora JH, Gillespie CE. M. A. (2024). \u003cem\u003eSystemic inflammation and lymphocyte activation precede rheumatoid arthritis\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2024.10.25.620344\u003c/span\u003e\u003cspan address=\"10.1101/2024.10.25.620344\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCho A, Ahn J, Kim A, Lee YJ, Song YW, Tanaka Y, Yi EC. A multi-biomarker panel for predicting Tocilizumab response in Rheumatoid arthritis patients. Translational Res. 2024;273:23\u0026ndash;31. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.trsl.2024.07.001\u003c/span\u003e\u003cspan address=\"10.1016/j.trsl.2024.07.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang Y, Hu Z, Huang R, Ho K, Wang P, Kang J. Metabolic reprogramming and macrophage expansion define ACPA-negative rheumatoid arthritis: Insights from single-cell RNA sequencing. Front Immunol. 2025;15:1512483. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2024.1512483\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2024.1512483\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBanda NK, Mehta G, Ferreira VP, Cortes C, Pickering MC, Pangburn MK, Arend WP, Holers VM. Essential Role of Surface-Bound Complement Factor H in Controlling Immune Complex\u0026ndash;Induced Arthritis. J Immunol. 2013;190(7):3560\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4049/jimmunol.1203271\u003c/span\u003e\u003cspan address=\"10.4049/jimmunol.1203271\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHietala MA, Jonsson I-M, Tarkowski A, Kleinau S, Pekna M. Complement Deficiency Ameliorates Collagen-Induced Arthritis in Mice. J Immunol. 2002;169(1):454\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4049/jimmunol.169.1.454\u003c/span\u003e\u003cspan address=\"10.4049/jimmunol.169.1.454\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHietala MA, Nandakumar KS, Persson L, Fahl\u0026eacute;n S, Holmdahl R, Pekna M. Complement activation by both classical and alternative pathways is critical for the effector phase of arthritis. Eur J Immunol. 2004;34(4):1208\u0026ndash;16. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/eji.200424895\u003c/span\u003e\u003cspan address=\"10.1002/eji.200424895\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBanda NK, Levitt B, Wood AK, Takahashi K, Stahl GL, Holers VM, Arend WP. Complement activation pathways in murine immune complex-induced arthritis and in C3a and C5a generation \u003cem\u003ein vitro\u003c/em\u003e. Clin Exp Immunol. 2009;159(1):100\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1365-2249.2009.04035.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1365-2249.2009.04035.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBanda NK, Takahashi M, Levitt B, Glogowska M, Nicholas J, Takahashi K, Stahl GL, Fujita T, Arend WP, Holers VM. Essential Role of Complement Mannose-Binding Lectin-Associated Serine Proteases-1/3 in the Murine Collagen Antibody-Induced Model of Inflammatory Arthritis. J Immunol. 2010;185(9):5598\u0026ndash;606. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4049/jimmunol.1001564\u003c/span\u003e\u003cspan address=\"10.4049/jimmunol.1001564\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSato Y, Sato R, Watanabe H, Kogure A, Watanabe K, Nishimaki T, Kasukawa R, Kuraya M, Fujita T. Complement activating properties of monoreactive and polyreactive IgM rheumatoid factors. Ann Rheum Dis. 1993;52(11):795\u0026ndash;800. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/ard.52.11.795\u003c/span\u003e\u003cspan address=\"10.1136/ard.52.11.795\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSabharwal UK, Vaughan JH, Fong S, Bennett PH, Carson DA, Curd JG. Activation of the classical pathway of complement by rheumatoid factors. Arthr Rhuem. 1982;25(2):161\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/art.1780250208\u003c/span\u003e\u003cspan address=\"10.1002/art.1780250208\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJahn B, Von Kempis J, Kr\u0026auml;mer KL, Filsinger S, H\u0026auml;nsch GM. Interaction of the terminal complement components C5b-9 with synovial fibroblasts: Binding to the membrane surface leads to increased levels in collagenase-specific mRNA. Immunology. 1993;78(2):329\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCrockard AD, Thompson JM, McBride SJ, Edgar JD, McNeill TA, Bell AL. Markers of inflammatory activation: Upregulation of complement receptors CR1 and CR3 on synovial fluid neutrophils from patients with inflammatory joint disease. Clin Immunol Immunopathol. 1992;65(2):135\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/0090-1229(92)90216-B\u003c/span\u003e\u003cspan address=\"10.1016/0090-1229(92)90216-B\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMerkel PA, George MD, Yue H, Popov T, Kronbichler A, Little MA, Jayne DRW. Safety of Avacopan for the Treatment of Antineutrophil Cytoplasmic Antibody\u0026ndash;Associated Vasculitis: Combined Data From Three Clinical Trials. ACR Open Rheumatol. 2025;7(4):e70001. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/acr2.70001\u003c/span\u003e\u003cspan address=\"10.1002/acr2.70001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrouw LA, Pickering MC, Blom AM. The complement system as a potential therapeutic target in rheumatic disease. Nat Rev Rheumatol. 2017;13(9):538\u0026ndash;47. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/nrrheum.2017.125\u003c/span\u003e\u003cspan address=\"10.1038/nrrheum.2017.125\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStorey JD. A Direct Approach to False Discovery Rates. J Royal Stat Soc Ser B: Stat Methodol. 2002;64(3):479\u0026ndash;98. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/1467-9868.00346\u003c/span\u003e\u003cspan address=\"10.1111/1467-9868.00346\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePope JE, Choy EH. C-reactive protein and implications in rheumatoid arthritis and associated comorbidities. Semin Arthritis Rheum. 2021;51(1):219\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.semarthrit.2020.11.005\u003c/span\u003e\u003cspan address=\"10.1016/j.semarthrit.2020.11.005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChambers RE, MacFarlane DG, Whicher JT, Dieppe PA. Serum amyloid-A protein concentration in rheumatoid arthritis and its role in monitoring disease activity. Ann Rheum Dis. 1983;42(6):665\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1136/ard.42.6.665\u003c/span\u003e\u003cspan address=\"10.1136/ard.42.6.665\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eConnolly M, Mullan RH, Mccormick J, Matthews C, Sullivan O, Kennedy A, FitzGerald O, Poole AR, Bresnihan B, Veale DJ, Fearon U. Acute-phase serum amyloid A regulates tumor necrosis factor α and matrix turnover and predicts disease progression in patients with inflammatory arthritis before and after biologic therapy. Arthr Rhuem. 2012;64(4):1035\u0026ndash;45. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/art.33455\u003c/span\u003e\u003cspan address=\"10.1002/art.33455\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSorić Hosman I, Kos I, Lamot L. Serum Amyloid A in Inflammatory Rheumatic Diseases: A Compendious Review of a Renowned Biomarker. Front Immunol. 2021;11:631299. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/fimmu.2020.631299\u003c/span\u003e\u003cspan address=\"10.3389/fimmu.2020.631299\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa MHY, Defranoux N, Li W, Sasso EH, Ibrahim F, Scott DL, Cope AP. A multi-biomarker disease activity score can predict sustained remission in rheumatoid arthritis. Arthritis Res Therapy. 2020;22(1):158. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s13075-020-02240-w\u003c/span\u003e\u003cspan address=\"10.1186/s13075-020-02240-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNaka T, Fujimoto M. LRG is a novel inflammatory marker clinically useful for the evaluation of disease activity in rheumatoid arthritis and inflammatory bowel disease. Immunological Med. 2018;41(2):62\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/13497413.2018.1481582\u003c/span\u003e\u003cspan address=\"10.1080/13497413.2018.1481582\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZou L, Fan Q, Liu Y, He H, Jia C. Leucine-rich alpha-2-glycoprotein 1 (LRG1) decrement during biologics therapy and its correlation with disease features and treatment outcomes in rheumatoid arthritis patients. Clin Rheumatol. 2025;44(9):3391\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10067-025-07545-2\u003c/span\u003e\u003cspan address=\"10.1007/s10067-025-07545-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQin Q, Song R, Du P, Gao C, Yao Q, Zhang J-A. (2021). Systemic Proteomic Analysis Reveals Distinct Exosomal Protein Profiles in Rheumatoid Arthritis. \u003cem\u003eJournal of Immunology Research\u003c/em\u003e, \u003cem\u003e2021\u003c/em\u003e, 9421720. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1155/2021/9421720\u003c/span\u003e\u003cspan address=\"10.1155/2021/9421720\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Q, Feng D, Song Y, Hu Z, Lu Q, Zhao M. Biased Usage of V/D/J Genes and Clonal Diversity in IgG Repertoires Correlates with Disease Activity and Clinical Features in Systemic Autoimmune Diseases. Immunol Investig. 2025;54(8):1461\u0026ndash;81. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/08820139.2025.2550374\u003c/span\u003e\u003cspan address=\"10.1080/08820139.2025.2550374\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLee PY, Osman J, Low TY, Jamal R. Plasma/Serum Proteomics: Depletion Strategies for Reducing High-Abundance Proteins for Biomarker Discovery. Bioanalysis. 2019;11(19):1799\u0026ndash;812. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4155/bio-2019-0145\u003c/span\u003e\u003cspan address=\"10.4155/bio-2019-0145\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables:","content":"\u003cdiv class=\"gridtable\"\u003e\u003cstrong\u003eTable 1: Clinical and demographic description of patients from the study cohort.\u003c/strong\u003e\n \u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eSeronegative\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003eSeropositive\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eDemographic Features\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eIndividuals, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e59\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eFemale, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e69% (33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e59% (35)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eAge, years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e63.9 (54.1\u0026ndash;72.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e59.5 (51.5\u0026ndash;67.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eEthnicity - White, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e100% (48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e97% (57)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eNon-Smoker, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e48% (23)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e36% (21)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSmoker, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e4% (2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e17% (10)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eFormer-Smoker, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e48% (23)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e46% (27)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eBaseline Clinical Features\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eBody Mass Index, kg/m\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e27.4 (24.4\u0026ndash;31.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e27.1 (23.6\u0026ndash;30.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eRF positive, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0% (0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3% (2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eACPA positive, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0% (0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e19% (11)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eACPA/RF positive, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0% (0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e78% (46)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSymptom duration, months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e3.4 (1.9-6)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e4.1 (2.8\u0026ndash;6.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eRheumatic disease comorbidity index (RDCI)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1 (0\u0026ndash;2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1 (0\u0026ndash;2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eCRP, mg/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e11.8 (4.3\u0026ndash;27.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e11.6 (4.1\u0026ndash;30.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eESR, mm/h\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e21 (13\u0026ndash;36)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e30.5 (20-48.8)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eBaseline Disease Activity Measures\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSwollen joint count (68)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e10 (8\u0026ndash;14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e13 (8\u0026ndash;21)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTender joint count (66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e12 (6\u0026ndash;18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e15 (10\u0026ndash;20)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDAS28-CRP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e5.3 (4.1-6)\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e5.2 (4.1\u0026ndash;6.2)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSDAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e31.9 (23.9\u0026ndash;40.7)\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e34.6 (22.6\u0026ndash;45.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eCDAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e30.8 (21.8\u0026ndash;38.6)\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e33.8 (20.4\u0026ndash;42.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eHAQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.9 (0.2\u0026ndash;1.1)\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1.1 (0.6\u0026ndash;1.5)\u003csup\u003ee\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePGA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e62 (49\u0026ndash;84)\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e57.5 (34.5\u0026ndash;77.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eEGA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e5 (2.9\u0026ndash;6.7)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e5.8 (3.5\u0026ndash;7.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"3\" nameend=\"c3\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eTreatment and Outcomes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDMARD exposure, months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e7 (5.2\u0026ndash;8.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e6.8 (4.9\u0026ndash;8.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePrednisone use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e17% (8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e12% (7)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eMTX alone use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e33% (16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e34% (20)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eMTX combination use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e40% (19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e53% (31)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ebDMARD use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e10% (5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e5% (3)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePatient in remission at 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e48% (23)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e54% (32)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eErosive patients at inclusion, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e10% (5)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e17% (10)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eErosive patients at 30m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e15% (7)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e37% (22)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSvH\u0026thinsp;\u0026gt;\u0026thinsp;=\u0026thinsp;5 patients, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e42% (20)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e61% (36)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eFootnote: Variables are presented as either median (IQR; 25th-75th) or % (n) without imputation of missing values. The MTX combination category comprises patients treated with MTX in combination with at least one additional DMARD at inclusion, or who escalated to an additional DMARD within 12 months. \u0026nbsp;(a: n = 58; b: n = 47; c: ; n = 45; d: n = 44; e: n = 56)\u003c/p\u003e\n\u003cp\u003eAnti-Cyclic Citrullinated Peptide Antibody (ACPA), Rheumatoid Factor (RF), C-Reactive Protein (CRP), Erythrocyte Sedimentation Rate (ESR), Disease Activity Score in 28 joints with CRP (DAS28-CRP), Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), \u0026nbsp; Health Assessment Questionnaire (HAQ), Patient Global Assessment (PGA), Evaluator Global Assessment (EGA), Sharp-van der Heijde total score (SvH).\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2: Patient characteristics by clusters based on baseline vs. 12-month proteome.\u003c/strong\u003e\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\n \u003cp\u003eBaseline\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e\n \u003cp\u003ePost-treatment\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eB1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003eB2\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003eP value\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003eP1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003eP2\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003eP value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eDemographic Features\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eIndividuals, n\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003eNA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eGender - Females, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e57% (28)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e69% (40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.287\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e50% (14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e68% (54)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.132\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eAge, years\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e59 (48.5\u0026ndash;67.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e62 (53.6\u0026ndash;69.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.328\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e61.2 (53.4\u0026ndash;69.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e59.5 (51.5\u0026ndash;68.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.840\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eBody Mass Index, kg/m\u0026sup2;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e27.3 (24.4\u0026ndash;29.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e27.2 (23.4\u0026ndash;31.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.571\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e27.1 (23.6\u0026ndash;30.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e27.4 (23.8\u0026ndash;31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.642\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eBaseline Clinical Features\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eACPA or RF positive, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e69% (34) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e43% (25) *\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.011\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e61% (17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e53% (42)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.639\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSymptom duration, months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e4.1 (2.3\u0026ndash;6.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3.6 (2.1-6)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.434\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e4.8 (2.6\u0026ndash;6.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e3.9 (2.1\u0026ndash;6.1)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.211\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSwollen joint count (68)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e11 (7\u0026ndash;18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e12 (8\u0026ndash;20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.389\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e10 (7\u0026ndash;17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e12 (8\u0026ndash;19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.558\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTender joint count (66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e13 (7\u0026ndash;18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e15 (7\u0026ndash;20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.386\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e14 (8\u0026ndash;17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e14 (7\u0026ndash;19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.739\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eCRP, mg/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e11.7 (5.6\u0026ndash;27.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e11.9 (3.6\u0026ndash;29.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.970\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e10.2 (3.1\u0026ndash;27.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e12.7 (5.5\u0026ndash;28.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.412\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eESR, mm/h\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e30 (18\u0026ndash;42)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e23 (16-46.5)\u003csup\u003ec\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.414\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e30 (16-46.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e24 (16\u0026ndash;42)\u003csup\u003ed\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.980\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eBaseline Disease Activity Measures\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDAS28-CRP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e5 (4\u0026ndash;6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e5.5 (4.6\u0026ndash;6.2)\u003csup\u003ee\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.211\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e5.1 (4.1\u0026ndash;6.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e5.3 (4.2\u0026ndash;6.1)\u003csup\u003ef\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.692\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSDAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e30.2 (22.5\u0026ndash;41.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e37.8 (24.5\u0026ndash;46.1)\u003csup\u003eg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.190\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e30.2 (22.4\u0026ndash;44.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e34.4 (23-44.9)\u003csup\u003eh\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.725\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eCDAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e28.4 (19.9\u0026ndash;38.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e33.9 (22\u0026ndash;40)\u003csup\u003eg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.198\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e29.9 (22.3\u0026ndash;41.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e32.6 (20.9\u0026ndash;39)\u003csup\u003eh\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.964\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eHAQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1 (0.6\u0026ndash;1.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.9 (0.4\u0026ndash;1.5)\u003csup\u003eg\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.881\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0.9 (0.5\u0026ndash;1.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e1 (0.5\u0026ndash;1.4)\u003csup\u003ei\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.912\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003e12-month Clinical Features\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSwollen joint count (68)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e2 (0\u0026ndash;8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2 (0\u0026ndash;6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.610\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e2 (0\u0026ndash;8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e2 (0\u0026ndash;6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.370\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTender joint count (66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e2 (0\u0026ndash;6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2 (0\u0026ndash;6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.814\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e2 (0\u0026ndash;6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e2 (0\u0026ndash;6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.539\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eCRP, mg/L\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1 (1\u0026ndash;8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1 (1-6.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.765\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e1 (1\u0026ndash;4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e1 (1-8.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.110\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eESR, mm/h\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e16 (6\u0026ndash;20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e14 (10\u0026ndash;22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.972\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e10.5 (8-18.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e16 (9.5\u0026ndash;22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.398\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003e12-month Disease Activity Measures\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDAS28-CRP\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e3.2 (1.7-4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2.5 (1.9\u0026ndash;3.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.819\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e2.5 (1.9\u0026ndash;3.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e2.6 (1.8-4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.935\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSDAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e10.1 (3.5\u0026ndash;18.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e8.1 (4\u0026ndash;14)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.626\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e9.4 (4.2\u0026ndash;19.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e9.5 (3.6\u0026ndash;15)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.606\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eCDAI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e9.8 (3.4\u0026ndash;17.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e7.8 (3.7\u0026ndash;13.9)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.660\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e8.9 (4.1\u0026ndash;19.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e8.6 (3.4\u0026ndash;14.5)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.469\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eHAQ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.1 (0-0.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.2 (0-0.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.248\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0.2 (0-0.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e0.2 (0-0.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.647\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"7\" nameend=\"c7\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eTreatment and Outcomes\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eDMARD exposure, months\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e7 (5.3\u0026ndash;8.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e6.8 (5.2\u0026ndash;8.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.802\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e6.8 (4.9\u0026ndash;8.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e7.2 (5.2\u0026ndash;8.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.530\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePrednisone use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e8% (4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e19% (11)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.162\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e7% (2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e16% (13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.344\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eMTX alone use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e39% (19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e29% (17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.408\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e32% (9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e34% (27)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eMTX combination use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e47% (23)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e47% (27)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e46% (13)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e47% (37)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e1.000\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ebDMARD use within 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e4% (2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e10% (6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.288\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0% (0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e10% (8)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.107\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePatient in remission at 12m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e49% (24)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e53% (31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.790\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e54% (15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e51% (40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.962\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eErosive patients at 30m, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e29% (14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e26% (15)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.967\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e21% (6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e29% (23)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.566\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSvH\u0026thinsp;\u0026gt;\u0026thinsp;=\u0026thinsp;5 patients, % (n)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e57% (28)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e48% (28)\u003csup\u003ea\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.529\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e50% (14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e53% (42)\u003csup\u003eb\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c7\"\u003e\n \u003cp\u003e0.897\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eFootnote: Variables are presented as either median (IQR; 25th-75th) or % (n) without imputation of missing values. The Mann-Whitney U test was used for continuous variables, and Chi-square or Fisher\u0026rsquo;s exact tests were applied for categorical variables. The MTX combination category comprises patients treated with MTX in combination with at least one additional DMARD at inclusion, or who escalated to an additional DMARD within 12 months. * Indicate statistical significance at p\u0026nbsp;\u0026le;0.05. (a: n = 57; b: n = 78; c: n = 56; d: n = 77; e: n = 54; f: n = 75; g: n = 53; h: n = 74; i: n = 73)\u003c/p\u003e\n\u003cp\u003eAnti-Cyclic Citrullinated Peptide Antibody (ACPA), Rheumatoid Factor (RF), C-Reactive Protein (CRP), Erythrocyte Sedimentation Rate (ESR), Disease Activity Score in 28 joints with CRP (DAS28-CRP), Simplified Disease Activity Index (SDAI), Clinical Disease Activity Index (CDAI), Health Assessment Questionnaire (HAQ).\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"arthritis-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"arrt","sideBox":"Learn more about [Arthritis Research \u0026 Therapy](http://arthritis-research.biomedcentral.com/)","snPcode":"13075","submissionUrl":"https://submission.nature.com/new-submission/13075/3","title":"Arthritis Research \u0026 Therapy","twitterHandle":"@ArthritisRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Mass Spectrometry, Rheumatoid Arthritis, Proteomics, RA biomarker, Erosion","lastPublishedDoi":"10.21203/rs.3.rs-9359440/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9359440/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eObjectives\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe aim to identify novel biomarkers of rheumatoid arthritis (RA) outcomes by performing serum proteomic profiling of our longitudinal RA cohort using Data-Independant Acquisition (DIA) mass spectrometry.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSerum proteomes from 107 previously untreated early RA patients were recruited in the EUPA cohort and their sera were analyzed at baseline and at the 12-month follow-up visit using DIA mass spectrometry technology. Clustering analyses on both baseline and follow-up serum profiles was performed to assess overall patient heterogeneity. Generalized estimating equations (GEE) analyses of longitudinal serum proteome data was used to query for proteins associated with disease activity or erosion outcomes. Functional networks of the identified proteins were explored using KEGG enrichment analysis while novel predictors of RA outcomes were identified using generalized linear models (GLM) with baseline serum proteomes, and performance was evaluated by receiver operating characteristic (ROC) curves.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRA patients could be separated in 2 distinct clusters based on their serum proteome profiles, regardless of disease activity or erosion outcomes. Compared to CRP, the protein signature composed of APOC4 and SAA1/SAA2 exhibited improved predictive performance for disease activity. Furthermore, an additional protein signature combining CTBS and IGHV5-10-1/IGHV5-51 outperformed classic autoantibodies serology status to predict erosiveness. Functional network enrichment analyses uncovered association between the complement system and RA progression.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLeveraging serum proteomic profiling, we identified novel biomarkers of RA outcomes, reinforcing the notion that protein signatures can improve predictive performance while highlighting crucial elements of pathophysiology that might lie in the complement system. Further clinical validation of these predictors may potentially pave the way toward improved personalized treatment strategies and ultimately better RA management.\u003c/p\u003e","manuscriptTitle":"Serum mass spectrometry unveil the heterogeneity of rheumatoid arthritis and reveals insights into Complement Pathways and IGHV as potential prognostic markers","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-12 10:35:50","doi":"10.21203/rs.3.rs-9359440/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"133148544282844392470003915281138219598","date":"2026-05-15T15:03:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"44455305706980015569199707753168410908","date":"2026-05-13T18:53:31+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-05-01T17:26:45+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-10T09:09:17+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-10T09:03:43+00:00","index":"","fulltext":""},{"type":"submitted","content":"Arthritis Research \u0026 Therapy","date":"2026-04-08T16:11:17+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"arthritis-research-and-therapy","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"arrt","sideBox":"Learn more about [Arthritis Research \u0026 Therapy](http://arthritis-research.biomedcentral.com/)","snPcode":"13075","submissionUrl":"https://submission.nature.com/new-submission/13075/3","title":"Arthritis Research \u0026 Therapy","twitterHandle":"@ArthritisRes","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"bb25344e-e892-4434-a4a9-e1c882ef6f6d","owner":[],"postedDate":"May 12th, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"133148544282844392470003915281138219598","date":"2026-05-15T15:03:41+00:00","index":46,"fulltext":""},{"type":"reviewerAgreed","content":"44455305706980015569199707753168410908","date":"2026-05-13T18:53:31+00:00","index":44,"fulltext":""},{"type":"reviewersInvited","content":"28","date":"2026-05-01T17:26:45+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-12T10:35:51+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-12 10:35:50","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9359440","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9359440","identity":"rs-9359440","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}