Related Experiment Video
Updated: May 30, 2026

Supervised Machine Learning for Semi-Quantification of Extracellular DNA in Glomerulonephritis
Published on: June 18, 2020
Re-evaluation of 129 patients with systemic necrotizing vasculitides by using classification algorithm according to
Sevil Kamali1, Bahar Artim-Esen, Burak Erer
1Division of Rheumatology, Department of Internal Medicine, Istanbul Medical Faculty, Istanbul University, Mevlana Cad. Topkapı Merkez Evleri 1. Etap A4 Blok Daire 13, Zeytinburnu, Istanbul, Turkey. sevilkamali@hotmail.com
Insights
This study validated a new classification algorithm for systemic necrotizing vasculitides. The algorithm proved helpful and practical, with most patients maintaining their original diagnoses, aiding epidemiological research.
Area of Science:
- Rheumatology
- Immunology
- Internal Medicine
Background:
- Systemic necrotizing vasculitides require accurate classification for diagnosis and research.
- Existing criteria, including the American College of Rheumatology (ACR) and Chapel Hill Consensus Criteria (CHCC), have limitations.
- A novel classification algorithm (CA) was proposed by Watts et al. to improve classification.
Purpose of the Study:
- To validate the newly proposed classification algorithm (CA) for systemic necrotizing vasculitides in a clinical cohort.
- To assess the agreement between the CA and existing diagnostic criteria.
- To evaluate the practicality and utility of the CA in patient reclassification.
Main Methods:
- Retrospective analysis of 129 patients with vasculitis.
- Reclassification of patients using the proposed CA, incorporating ACR, CHCC, and Sorensen (So) surrogate markers.
- Statistical analysis using Kappa statistic to assess agreement between criteria.
Main Results:
- The majority of patients (12 CSS, 91% WG, 78% MPA, 93% c-PAN) retained their prior diagnoses after applying the CA.
- Significant agreement was observed between ACR and So criteria in Wegener's granulomatosis (WG) (κ=0.62).
- The CA successfully reclassified some previously unclassified patients and facilitated categorization within WG and microscopic polyangiitis (MPA).
Conclusions:
- The CA is a helpful and practical tool for classifying systemic necrotizing vasculitides, particularly in epidemiological studies.
- The algorithm demonstrated good concordance with previous diagnoses for most patients.
- Discrepancies in classification may arise from the diverse features (clinical, histopathological, serological) captured by different criteria.
Abstract:
Recently, a new classification algorithm (CA) for systemic necrotizing vasculitides was proposed by Watts et al. (Annals Rheum Dis 66:222-227, 2007) by using the American College of Rheumatology (ACR), Chapel Hill Consensus Criteria (CHCC) and Sorensen surrogate markers (So). We aimed to validate CA in our patients. One hundred twenty-nine patients followed up in our vasculitis clinic were reclassified according to CA in different categories (ACR or Lanham criteria in "1" for Churg-Strauss Syndrome (CSS); ACR in "2a"; CHCC-Wegener's granulomatosis (WG) in "2b"; CHCC-microscopic polyangiitis (MPA), So-WG in "2c"; So-WG, proteinase 3 (PR3) or myeloperoxidase antineutrophil cytoplasmic antibody (MPO ANCA) serology in "2d" for WG; clinical features and histology compatible with small vessel vasculitis without So-WG in "3a"; So-MPA, PR3 or MPO ANCA serology in "3b" for MPA; CHCC-classic-polyarteritis nodosa (c-PAN) or typical angiographic features in "4" for c-PAN; unclassifiable in "5"). Kappa statistic was used to analyse the agreement of the criteria that formed the algorithm. All of 12 CSS, 91% of 69 WG, 78% of 18 MPA and 93% of 26 c-PAN patients remained in their previous diagnosis. WG patients were placed in 2a (83%), 2c (3%), 2d (14%) categories. Four WG (6%) and four MPA (22%) patients were categorized as MPA (in 3a (75%), 3b (25%)) and WG (in 2c (75%), 2d (25%)), respectively. Three of four unclassified patients could be classified as c-PAN (two) and MPA (one). Significant agreement was demonstrated only for ACR and So criteria in WG (κ = 0.62, p < 0.001). The majority of our patients stayed on their previous diagnosis in "CA". Our findings suggest that this algorithm is helpful and practical for epidemiological studies. Poor correlation of defined criteria was thought to be related to the fact that each criteria mainly consist of different characteristics of vasculitides such as clinical, histopathological and serological features.
Related Concept Videos
Nephrotic Syndrome II : Assessment and Medical Management
Acute Pyelonephritis II: Diagnostic Studies and Management
Heart Failure IV: Classification and Diagnostic Evaluation
Varicose Veins II: Diagnostic Studies and Interprofessional Care
