Using explainable artificial intelligence to predict and forestall flare in rheumatoid arthritis.
Using explainable artificial intelligence to predict and forestall flare in rheumatoid arthritis.
复制标题
使用可解释的人工智能来预测和预防类风湿关节炎的发作。
DOI：
10.1038/s41591-024-02818-w
复制
发表时间：
2024
期刊：
影响因子：
82.9
通讯作者：
Alivernini S
中科院分区：
文献类型：
--
作者：
Alivernini S
目前治疗类风湿性关节炎的目的是实现和维持疾病缓解，其特征在于没有炎症症状1。然而，缓解往往是脆弱的，约50%的患者在减少或停止药物治疗后会出现疾病发作，这对患者和临床医生都是一种负担。虽然存在一些复发的预测因子，但它们不能可靠地预测大多数缓解期类风湿关节炎患者的复发2。最近的研究表明，患者滑膜组织细胞和分子特征的解卷积可能提供更准确的预测模型3，4。根据临床症状和超声检查定义的持续缓解期类风湿关节炎患者可能存在残留的滑膜组织炎症，这可能决定缓解的可持续性5。单细胞转录组学的最新进展揭示了缓解期间类风湿性关节炎滑膜组织的先前未识别的异质性，并确定了预测爆发或持续缓解的滑膜巨噬细胞和成纤维细胞的不同簇4，6。例如，在治疗停止后持续缓解的类风湿性关节炎中，存在健康的Mer酪氨酸激酶（MerTK）+ CD206+滑膜组织巨噬细胞（STM）簇的恢复，所述簇是TREM2+和LYVE 1+并且具有炎症消退特性。相反，缓解期患者随后发作，具有持续性致病性MerTK− CD206− STM，具有CD48+S100A12+表型，这是活动性类风湿关节炎的特征4。这支持了对个体患者的滑膜组织的分子特征进行去卷积以用于开发预测模型的需要。
Current treatments for rheumatoid arthritis aim to achieve and maintain disease remission, which is characterized by the absence of symptoms of inflammation 1. However, remission is often fragile, with about 50% of patients experiencing disease flares after reducing or stopping medication, representing a burden for both patients and clinicians 1. Although some predictors of flare exist, they do not confidently predict flare for most patients with rheumatoid arthritis in remission 2. Recent studies suggest that deconvolution of patients’ synovial tissue cellular and molecular signatures may offer more accurate prediction models 3, 4.Patients with rheumatoid arthritis in sustained remission, defined on the basis of clinical symptoms and ultrasound examination, may have residual synovial tissue inflammation that could determine the sustainability of remission 5. Recent advances in single-cell transcriptomics have uncovered a previously unrecognized heterogeneity of rheumatoid arthritis synovial tissue during remission and identified distinct clusters of synovial macrophages and fibroblasts that are predictive of flare or sustained remission 4, 6. For example, in rheumatoid arthritis in remission that is sustained after treatment withdrawal, there is a restoration of healthy Mer tyrosine kinase (MerTK)+ CD206+ synovial tissue macrophage (STM) clusters that are TREM2+ and LYVE1+ and have inflammationresolving properties. Conversely, patients in remission who subsequently flare have persistent pathogenic MerTK− CD206− STMs with a CD48+S100A12+ phenotype, which is characteristic of active rheumatoid arthritis 4. This supports the need for deconvolution of the molecular signatures of the synovial tissue of individual patients for the development of prediction models.
影响因子：
27.4
作者：
Alivernini S;Tolusso B;Petricca L;Bui L;Di Sante G;Peluso G;Benvenuto R;Fedele AL;Federico F;Ferraccioli G;Gremese E
通讯作者：
Gremese E
影响因子：
12.8
作者：
Baker, Kenneth F.;Skelton, Andrew J.;Isaacs, John D.
通讯作者：
Isaacs, John D.