Vitte Joana, Ranque Stéphane, Carsin Ania, Gomez Carine, Romain Thomas, Cassagne Carole, Gouitaa Marion, Baravalle-Einaudi Mélisande, Bel Nathalie Stremler-Le, Reynaud-Gaubert Martine, Dubus Jean-Christophe, Mège Jean-Louis, Gaudart Jean
Aix-Marseille Univ, APHM Assistance Publique Hôpitaux de Marseille, Hôpital de La Conception, Laboratoire d'Immunologie, Marseille, France.
Aix-Marseille Univ, UMR INSERM 1067 CNRS 7333, Marseille, France.
Front Immunol. 2017 Aug 22;8:1019. doi: 10.3389/fimmu.2017.01019. eCollection 2017.
Molecular-based allergy diagnosis yields multiple biomarker datasets. The classical diagnostic score for allergic bronchopulmonary aspergillosis (ABPA), a severe disease usually occurring in asthmatic patients and people with cystic fibrosis, comprises succinct immunological criteria formulated in 1977: total IgE, anti- () IgE, anti- "precipitins," and anti- IgG. Progress achieved over the last four decades led to multiple IgE and IgG(4) biomarkers available with quantitative, standardized, molecular-level reports. These newly available biomarkers have not been included in the current diagnostic criteria, either individually or in algorithms, despite persistent underdiagnosis of ABPA. Large numbers of individual biomarkers may hinder their use in clinical practice. Conversely, multivariate analysis using new tools may bring about a better chance of less diagnostic mistakes. We report here a proof-of-concept work consisting of a three-step multivariate analysis of IgE, IgG, and IgG4 biomarkers through a combination of principal component analysis, hierarchical ascendant classification, and classification and regression tree multivariate analysis. The resulting diagnostic algorithms might show the way for novel criteria and improved diagnostic efficiency in -sensitized patients at risk for ABPA.
基于分子的过敏诊断可产生多个生物标志物数据集。过敏性支气管肺曲霉病（ABPA）是一种通常发生在哮喘患者和囊性纤维化患者身上的严重疾病，其经典诊断评分包括1977年制定的简洁免疫学标准：总IgE、抗（）IgE、抗“沉淀素”和抗IgG。过去四十年来取得的进展使得有多种IgE和IgG(4)生物标志物可提供定量、标准化的分子水平报告。尽管ABPA一直存在诊断不足的情况，但这些新出现的生物标志物无论是单独还是在算法中都未被纳入当前的诊断标准。大量的个体生物标志物可能会阻碍它们在临床实践中的应用。相反，使用新工具进行多变量分析可能会带来减少诊断错误的更好机会。我们在此报告一项概念验证工作，该工作通过主成分分析、层次上升分类和分类与回归树多变量分析相结合的方式，对IgE、IgG和IgG4生物标志物进行了三步多变量分析。所得出的诊断算法可能为ABPA风险致敏患者的新诊断标准和提高诊断效率指明方向。
