分析不同糖皮质激素(GC)反应性的急性淋巴细胞白血病(ALL)患儿初诊时的实验室检测结果差异，寻找可以预测GC反应性的指标并构建相应的预测模型，为早期评估患儿的GC反应性，优化化疗方案提供参考依据。

对首诊于中山大学附属第一医院的ALL患儿的病例资料进行回顾性分析，收集并统计其初诊时诱导化疗前的相关临床资料和实验室检测结果，筛选GC抵抗的相关高危因子，并利用二元logistic回归方程及列线图构建预测模型，使用ROC曲线评价其诊断效能。

初诊时白细胞计数(WBC)、活化部分凝血活酶时间(APTT)、凝血因子Ⅷ、K⁺、Cl^-、CO₂、LDH、BCR/ABL1、ETV6/RUNX1在GC敏感组和抵抗组之间差异有统计学意义。其中APTT、K⁺和WBC是GC反应性的独立危险预测因子，综合这三项血液学指标构建的预测模型其诊断敏感度为0.80，特异度为0.84。

本研究提示综合诱导化疗前APTT、K⁺和WBC三项实验室检测指标的预测模型能够更好地为临床判断ALL患儿的GC反应性提供参考，但仍然需要更多的病例数据进一步验证。

To provide reference for early assessment of GC reactivity in children and optimization of chemotherapy, the differences in laboratory test results of children with acute lymphoblastic leukemia (ALL) with different glucocorticoid (GC) reactivity at initial diagnosis were analyzed to find indicators that could predict GC reactivity and construct corresponding prediction models.

We retrospectively analyzed the cases of children with ALL who were first diagnosed by the First Affiliated Hospital of Sun Yat-sen University. The clinical data and laboratory test results before induction chemotherapy were collected and counted, and the relevant risk factors of GC-resistant were screened. The binary logistic regression equation and nomogram were used to construct the prediction model, and the ROC curve was used to evaluate the diagnostic efficiency.

The laboratory test results at the initial diagnosis such as WBC count, APTT, coagulation factor 8, K⁺, Cl^-, CO₂, LDH, BCR-ABL1, ETV6-RUNX1 showed significant statistical differences between the ALL GC-resistant group and GC-sensitive group. APTT, K⁺ and WBC count were independent risk predictors of GC reactivity. The sensitivity and specificity of the predictive model were 0.80 and 0.84, respectively.

This study suggested that the prediction model combining APTT, K⁺ and white blood cell count before induction chemotherapy could provide a better reference for clinical judgment of GC reactivity in children with ALL, but more cases are still needed for further verification.