Objective: To assess the value of serum tumor specific protein 70 (SP70) for prognostic stratification in acute myeloid leukemia (AML). Methods: A cohort study design was adopted. 129 newly diagnosed AML patients from September 2022 to January 2024 at the Hematology Department of the First Affiliated Hospital of Nanjing Medical University were included, as well as a control group consisted of 120 healthy individuals and 7 cases with benign hematologic diseases during the same period (total 127 cases). Clinical data were collected from Electronic Medical Records. According to the 2023 edition of the Chinese Leukemia Diagnosis and Treatment Guidelines, AML patients with good or moderate prognosis were categorized as low-to-intermediate risk, while those with poor prognosis were high-risk group. Univariate and multivariate logistic regression analyses were used to identify variables significantly associated with AML prognostic risk. ROC analysis was used to evaluate diagnostic performance. A nomogram for predicting patient prognostic risk was constructed by R 4.0.2 software, and the internal validation was performed using bootstrapping. Results: Among 129 AML patients, there were 71 males (55.0%) and 58 females (45.0%), with 42 (32.6%) classified as high-risk and 87 (67.4%) as low-intermediate risk. The high-risk group had a significantly higher median age [62 (48, 67) years] compared to the low-intermediate risk group [50 (35, 63) years, Z=-2.381, P=0.017], and a significantly higher proportion of males (30 patients, 71.4%) compared to the low-intermediate risk group (41 patients, 47.1%, χ2=6.760, P=0.009). Multivariate logistic regression analysis indicated that serum SP70 (OR=2.54, 95%CI: 1.68-3.84, P<0.001), hemoglobin (HB) (OR=0.96, 95%CI: 0.93-0.99, P<0.05), and bone marrow blast (BM blast) (OR=1.07, 95%CI: 1.02-1.13, P<0.05) were independent risk factors for high-risk prognosis in AML patients. ROC analysis showed that the area under the curve (AUC) for SP70 predicting high-risk patients was 0.908 (cut-off value of 5.74 ng/ml, 95%CI: 0.845-0.952, sensitivity 90.5%, specificity 82.8%). The combined model of serum SP70, HB, and BM blasts had an AUC of 0.931 (95%CI: 0.890-0.973); C-index=0.925 (95%CI: 0.876-0.963),with no statistically significant difference compared to serum SP70 alone (Z=1.693,P>0.05). Conclusion: Serum SP70 may be a promising non-invasive molecular biomarker for prognostic stratification in AML.
目的： 探讨血清肿瘤特异蛋白70（SP70）在急性髓系白血病（AML）预后分层评估中的应用价值。 方法： 采用队列研究设计，收集2022年9月至2024年1月于南京医科大学第一附属医院血液科初次确诊为急性髓系白血病患者129例，以同时期120名健康体检者和血液良性疾病患者7例作为对照（共127例）。通过病历系统收集患者临床资料，根据中国白血病诊疗指南（2023版）将AML预后良好与预后中等者归为低中危组、预后不良者归为高危组。采用单因素和多因素logistic回归分析，选择与AML预后分层显著相关的变量，绘制ROC曲线，并用R 4.0.2软件构建预测患者预后分层的列线图，以bootstrapping法进行模型内部验证。 结果： 129例AML患者中，男性71例（55.0%），女性58例（45.0%），其中高危组42例（32.6%），低中危组87例（67.4%）。高危组的年龄［62（48，67）岁］显著高于低中危组［50（35，63）岁，Z=-2.381，P=0.017］，其男性占比（30例，71.4%）显著高于低中危组（41例，47.1%，χ2=6.760，P=0.009）。多因素logistic回归分析结果提示血清SP70（OR=2.54，95%CI：1.68~3.84，P<0.001）、血红蛋白浓度（HB）（OR=0.96，95%CI：0.93~0.99，P<0.05）和骨髓原始细胞比例（BM blasts）（OR=1.07，95%CI：1.02~1.13，P<0.05）是患者预后高危的独立危险因素。ROC分析显示，血清SP70用于鉴别患者预后高危的曲线下面积（AUC）为0.908（cut-off值5.74 ng/ml，95%CI：0.845~0.952，灵敏度90.5%，特异度82.8%）。血清SP70、HB和BM blasts三者联合构建模型的AUC=0.931（95%CI：0.890~0.973）；C-index为0.925（95%CI：0.876~0.963），其与血清SP70单指标鉴别效能间差异无统计学意义（Z=1.693，P>0.05）。 结论： 血清SP70可能是一项用于AML预后分层有潜力的生物标志物。.