摘要目的:比较流式细胞术检测多聚二磷酸腺苷核糖聚合酶-1（PARP-1）和增殖指数Ki67在慢性淋巴细胞白血病（CLL）的临床价值，同时对其预后价值进行评估。方法:回顾性队列研究，纳入南京医科大学第一附属医院2015年1月1日至2024年9月13日147例初诊CLL患者，其中男94例，女53例，年龄60（52，70）岁。查阅患者的电子病历系统，收集患者或初诊时的临床基线资料。通过多参数流式细胞术（MFC）检测这147例CLL患者肿瘤细胞PARP-1和Ki67的表达。根据随访期间是否发生进展，定义2例转为aCLL和2例Richter转化CLL患者为进展期（4例），未发生进展的Binet C期为Binet C期 NP（65例）。使用Python随机分为训练集（103例）和验证集（44例），训练集根据是否存在治疗指征，将患者分为治疗组（73例）和随访组（30例）。采用Mann-Whitney检验和Kruskal-Walis检验比较不同临床及遗传学特征的CLL患者Ki67与PARP-1表达水平，受试者工作（ROC）曲线界定PARP-1的阈值，曲线下面积（AUC）的比较采用DeLong检验。Kaplan-Meier分析CLL患者的中位无治疗生存期（TFS）。 结果:在不同Binet分期（A期、B期和C期）中PARP-1表达率差异有统计学意义（4.22%比12.30%比27.80%， H=34.13， P<0.001），Binet C期 NP与进展期PARP-1表达率差异亦有统计学意义（27.40%比87.20%， U=38， P<0.05）；而Ki67表达率仅在Binet C期 NP与进展期之间差异有统计学意义（0%比8.41%， Z=17， P<0.001）。免疫球蛋白重链可变区（IGHV）无突变（28.5%比12.0%， U=982.5， P<0.001）、血红蛋白<110 g/L（32.65%比11.87%， U=1 331， P<0.001）、血小板<100×10 9/L（27.4%比10.5%， U=1 424， P<0.001）、β 2微球蛋白>3.5 mg/L（22.7%比11.94%， U=1 441， P<0.05）、脾大（25.3%比10.41%， U=1 640， P<0.001）的CLL患者PARP-1表达率较高，且差异有统计学意义。此外，PARP-1（AUC=0.802）的受试者工作特征曲线下面积要明显优于Ki67（AUC=0.512）（ Z=5.71， P<0.001）。Kaplan-Meier分析显示，PARP-1>12.15%的患者无治疗生存期（TFS）更差（ P<0.001）。 结论:MFC检测PARP-1是比Ki67更为理想的CLL增殖指数和预后标记物。

abstractsObjective:To compare the clinical value of Poly(ADP-ribose) polymerase-1 (PARP-1) and the proliferation index Ki-67 detected by flow cytometry in chronic lymphocytic leukemia (CLL), and to evaluate their prognostic value.Methods:A retrospective cohort study was conducted, including 147 newly diagnosed CLL patients at the First Affiliated Hospital of Nanjing Medical University from January 1, 2015, to September 13, 2024. The cohort consisted of 94 males and 53 females, with a median age of 60 (52,70) years. Baseline clinical characteristics at the time of initial diagnosis were retrieved from the electronic medical record system.The expression of PARP-1 and Ki-67 in tumor cells was detected by multiparametric flow cytometry (MFC). Based on disease progression during follow-up, patients were categorized into a progressive group [4 cases; including 2 cases of transformation to aggressive CLL (aCLL) and 2 cases of Richter′s transformation] and a Binet C non-progressive group (Binet C NP, 65 cases). Using Python, patients were randomly divided into a training set (103 cases) and a validation set (44 cases). The training set was further divided into a treatment group (73 cases) and a follow-up group (30 cases) based on the presence of treatment indications. Mann-Whitney U test and Kruskal-Wallis test were used to compare Ki-67 and PARP-1 expression levels across different clinical and genetic characteristics. Receiver Operating Characteristic (ROC) curves were used to determine the threshold for PARP-1.The comparison of the AUC was performed using DeLong′s test. Kaplan-Meier analysis was utilized to assess the median Treatment-Free Survival (TFS). Results:There were statistically significant differences in PARP-1 expression rates among different Binet stages (A, B, and C) (4.22% vs. 12.30% vs. 27.80%; H=34.13, P<0.001). Similarly, the difference in PARP-1 expression between the Binet stage C NP group and the progressive group was also significant (27.40% vs. 87.20%; U=38, P<0.05). In contrast, Ki67 expression rates only differed significantly between the Binet stage C NP group and the progressive group (0% vs. 8.41%; Z=17, P<0.001). Significantly higher PARP-1 expression rates were observed in CLL patients with unmutated IGHV (28.5% vs. 12.0%; U=982.5, P<0.001), hemoglobin<110 g/L (32.65% vs. 11.87%; U=1 331, P<0.001), platelet<100×10 9/L (27.4% vs. 10.5%; U=1 424, P<0.001), β 2?-microglobulin>3.5 mg/L (22.7% vs. 11.94%; U=1441, P<0.05), and splenomegaly (25.3% vs. 10.41%; U=1 640, P<0.001). Furthermore, the area under the ROC curve of PARP-1 (0.802) was significantly superior to that of Ki67 (0.512) ( Z=5.71, P<0.001).Kaplan-Meier analysis indicated that patients with PARP-1>12.15% had significantly worse treatment-free survival (TFS) ( P<0.001). Conclusions:MFC-detected PARP-1 is a more ideal proliferation index and prognostic marker for CLL compared to Ki-67.