Unveiling the Unseen Signals: Identifying Novel Diagnostic and Prognostic Biomarkers for Eosinophilic Granulomatosis with Polyangiitis

Background Eosinophilic granulomatosis with polyangiitis (EGPA) is a rare systemic vasculitis characterized by necrotizing inflammation of small to medium-sized blood vessels. EGPA is linked to asthma, chronic sinusitis with nasal polyps, and a notable increase in eosinophils in tissue and blood. However, the disease displays considerable heterogeneity in its clinical manifestations. Despite noteworthy advancements in the classification and treatment of EGPA, persistent challenges persist in terms of diagnosis, prognosis and stratification, primarily attributable to the absence of specific diagnostic biomarkers. 1.2 Objectives The study was conducted with the objective of addressing critical gaps in the diagnosis and prognosis of EGPA by means of evaluating novel biomarkers and eosinophil morphological characteristics. Firstly, the morphological analysis of eosinophils in active EGPA was conducted to assess potential dysplasia and its diagnostic value. Secondly, the investigation of anti-pentraxin 3 (anti-PTX3) antibodies as a potential diagnostic marker was undertaken. Thirdly, rheumatoid factor (RF) was evaluated as a prognostic marker for disease relapses. By exploring these parameters, this study would enhance the clinical management of EGPA by improving diagnostic accuracy and predicting disease relapse more effectively. 1.3 Methods A multi-phase study design was implemented, encompassing cross-sectional and retrospective cohort studies. The eosinophil morphology study included EGPA patients with active disease and hypereosinophilia, compared against a control group with drug-induced eosinophilia. Cytological evaluation was conducted using MGG staining, with morphological classification adhering to established guidelines. The anti-PTX3 antibody study involved a cohort of EGPA patients alongside control groups of severe eosinophilic asthma (SEA) and chronic rhinosinusitis with nasal polyposis (CRSwNP) patients, with serum analysis performed using an homemade ELISA. The RF cohort study retrospectively analyzed EGPA patients stratified by RF status, employing machine learning techniques for relapse prediction, including Random Forest. 1.4 Results The eosinophil morphology assessment demonstrated a tendency towards a higher prevalence of eosinophilic dysplasia in EGPA in comparison to drug-induced eosinophilia. The anti-PTX3 study revealed a significantly higher positivity rate (p<0.0001) for anti-PTX3 in EGPA (36.2%) in comparison to SEA (1.7%) and CRSwNP (4.8%), supporting its potential as a diagnostic marker, particularly for ANCA-negative cases. No significant differences were observed between anti-PTX3-positive and anti-PTX3-negative EGPA patients concerning clinical and prognostic parameters at diagnosis. The RF cohort study identified RF positivity in 45.8 % of EGPA patients, with RF-positive patients demonstrating a significantly higher relapse rate rather than RF-negative patients (39.5% vs. 2.2%, p=0.003). Machine learning analysis confirmed RF as the most significant predictor of relapse, with high levels of both sensitivity (96.3%) and specificity (89.3%) in Random Forest models. Kaplan-Meier analysis of the five-year relapse rate further supported the association between RF positivity and shorter 5 years relapse-free survival (log-rank test p<0.001). Furthermore, SHAP values demonstrated that RF positivity contributed the most to relapse prediction, with a phi value of +0.1469. Conclusion This study underscores the importance of novel biomarkers in enhancing EGPA diagnosis and prognosis. Anti-PTX3 antibodies emerged as a promising diagnostic tool, particularly for ANCA-negative EGPA, though further validation is required. RF was identified as a robust prognostic marker for relapse, providing a valuable stratification tool for clinical management. The findings advocate for integrating these biomarkers into routine practice to improve early diagnosis and personalized treatment approaches.