# Integrating clinical features, inflammatory markers, and immune profiles: a Yunke-based nomogram model for rheumatoid arthritis prognosis

## TL;DR

This study creates a predictive model to help doctors tailor treatment for rheumatoid arthritis patients using clinical data, inflammation markers, and immune profiles.

## Contribution

The first nomogram integrating clinical, inflammatory, and immune parameters for prognosis in Yunke-combined therapy for rheumatoid arthritis.

## Key findings

- Six independent predictors were identified, including rheumatoid factor, CRP, swollen joints, TNF-α, IL-6, and CD3+ T cells.
- The nomogram showed strong discrimination with C-indexes of 0.883 in training and 0.823 in validation cohorts.
- The model achieved high sensitivity and specificity in predicting treatment response and supports personalized therapeutic monitoring.

## Abstract

To develop a prognostic nomogram integrating clinical, inflammatory, and immune parameters for rheumatoid arthritis (RA) patients receiving Yunke-drug combination therapy, facilitating personalized treatment decisions.

We retrospectively analyzed 304 RA patients (2010–2024) divided into training (n = 213) and validation (n = 91) cohorts. Predictor selection through univariate/multivariate logistic regression informed nomogram construction. Model performance was assessed via ROC curves, calibration plots, and decision curve analysis (DCA).

Six independent predictors emerged: elevated rheumatoid factor (OR = 1.32, 1.08–1.62), CRP > 10 mg/L (OR = 2.14, 1.45–3.16), ≥4 swollen joints (OR = 1.87, 1.22–2.88), TNF-α > 8.1 pg./mL (OR = 2.05, 1.33–3.17), IL-6 > 15 pg./mL (OR = 1.94, 1.25–3.01), and CD3 + T cells <650/μL (OR = 1.76, 1.15–2.70) (all p < 0.05). The nomogram showed strong discrimination (C-index: 0.883 training; 0.823 validation) with AUCs of 0.881 (0.804–0.958) and 0.823 (0.679–0.966). Sensitivity/specificity reached 94.3%/90.7% (training) versus 78.3%/81.2% (validation). DCA confirmed clinical utility across probability thresholds (15–85%).

This first multifactorial nomogram for Yunke-combined therapy integrates joint assessments, serum biomarkers, cytokine profiles, and cellular immunity indicators. Demonstrated predictive accuracy (30.5% training; 29.7% validation response rates) supports its potential for therapeutic monitoring. While internally validated, multicenter studies are required to confirm generalizability. The model establishes a framework for precision RA management, with implications for dose optimization and resistance mechanism research.