医学
内科学
髓系白血病
列线图
接收机工作特性
纤维蛋白原
胃肠病学
递归分区
多元分析
肿瘤科
外科
作者
Yaqun Ding,Xiangyu Qi,Li Yang,Yanni Sun,Jia Wan,Chengxin Luo,Yarui Huang,Qingrong Li,Guixian Wu,Xiaoqing Zhu,Shuangnian Xu
标识
DOI:10.1007/s10238-023-01241-8
摘要
Inflammation and nutrition related proteins participate in the development of acute myeloid leukemia (AML). It has been reported that the albumin-to-fibrinogen ratio (AFR) could serve as a prognostic indicator in patients with malignancy, but the precise relevance of AML is unclear. This study aimed to evaluate the effect of AFR on survival prognosis in patients with AML. We analyzed 227 patients newly diagnosed with non-M3 AML. AFR was calculated as albumin divided by fibrinogen. Based on the cutoff point from X-tile program, patients were divided into AFR-high (38.8%) and AFR-low (61.2%) groups. AFR-low group showed a poorer complete remission rate (P < 0.001) and median time to relapse (P = 0.026), while the mortality was higher (P = 0.009) than AFR-high ones. According to the log-rank test, AFR-low group had shorter OS (P < 0.001) and DFS (P = 0.034). Multivariate analysis identified AFR, ELN risk, bone marrow transplant, and hemoglobin as independent prognostic variables associated with OS. A visualized nomogram for predicting OS was performed. The C-index (0.75), calibration plots, and decision curve analyses of new model showed better discrimination, calibration, and net benefits than the ELN risk model. The time-dependent receiver operating characteristic (ROC) curve of 1-, 2-, and 3-year also functioned well (AUC, 0.81, 0.93 and 0.90, respectively). Our study provided a comprehensive view of AFR which could be an independent prognostic indicator in AML patients. The prognostic model utilized readily available information from ordinary clinical practice to improve predictive performance, identify risks, and assist in therapeutic decision-making.
科研通智能强力驱动
Strongly Powered by AbleSci AI