原发性免疫性血小板减少症脾切除远期应答的蛋白质组学图谱与机器学习预测
研究回顾性纳入34例、前瞻性纳入26例原发性免疫性血小板减少症患者,于脾切除前采集骨髓活检样本行蛋白质组学分析。加权基因共表达网络分析识别12个蛋白模块,机器学习在训练队列从蓝色模块筛出生物标志物组合,丝切蛋白-1(CFL1)在验证队列获独立确认,两队列C指数均>0.7。
要点
- 脾切除前骨髓蛋白质组学可用于筛选预测脾切除应答的关键分析物。
- 蓝色模块与脾切除应答显著相关(r=0.37,p=0.0031)。
- 丝切蛋白-1(CFL1)在训练与验证队列中C指数均>0.7。
核心结论
- 骨髓蛋白质组学分析可用于筛选预测原发性免疫性血小板减少症患者脾切除应答的关键分析物。
- 作者认为需在该患者人群中进一步探索血浆蛋白质组学。
研究设计与样本
- 回顾性训练队列纳入34例原发性免疫性血小板减少症患者,前瞻性验证队列纳入26例。
- 全部参与者在脾切除前采集骨髓活检样本。
- 采用加权基因共表达网络分析(WGCNA)在开发队列中识别蛋白模块。
主要发现
- 共识别12个蛋白模块。
- tan模块与脾切除前巨核细胞计数正相关(r=0.38,p=0.027)。
- tan模块与脾切除后血小板峰值时间正相关(r=0.47,p=0.005)。
- 蓝色模块与脾切除应答显著相关(r=0.37,p=0.0031)。
- KEGG通路分析显示tan模块主要富集PI3K-Akt信号通路,蓝色模块富集核糖体与剪接体通路。
关键数值与推荐
- 机器学习算法在训练队列从蓝色模块识别出最佳生物标志物组合。
- 丝切蛋白-1(CFL1)在验证队列获独立确认,两队列C指数均>0.7。
临床怎么用
- 原文未据此给出临床推荐。
局限性
- 作者认为需在该患者人群中进一步探索血浆蛋白质组学。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
This study aimed to identify key proteomic analytes correlated with response to splenectomy in primary immune thrombocytopenia (ITP). Thirty-four patients were retrospectively collected in the training cohort and 26 were prospectively enrolled as validation cohort. Bone marrow biopsy samples of all participants were collected prior to the splenectomy. A total of 12 modules of proteins were identified by weighted gene co-expression network analysis (WGCNA) method in the developed cohort. The tan module positively correlated with megakaryocyte counts before splenectomy (r = 0.38, p = 0.027), and time to peak platelet level after splenectomy (r = 0.47, p = 0.005). The blue module significantly correlated with response to splenectomy (r = 0.37, p = 0.0031). KEGG pathways analysis found that the PI3K-Akt signalling pathway was predominantly enriched in the tan module, while ribosomal and spliceosome pathways were enriched in the blue module. Machine learning algorithm identified the optimal combination of biomarkers from the blue module in the training cohort, and importantly, cofilin-1 (CFL1) was independently confirmed in the validation cohort. The C-index of CFL1 was >0.7 in both cohorts. Our results highlight the use of bone marrow proteomics analysis for deriving key analytes that predict the response to splenectomy, warranting further exploration of plasma proteomics in this patient population.
引用
- Sun T, Chen J, Xu Y, et al. Proteomics landscape and machine learning prediction of long-term response to splenectomy in primary immune thrombocytopenia. Br J Haematol. 2024. PMID: 38513635. doi:10.1111/bjh.19420
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。