电化学发光法检测尿Cyfra21-1对膀胱癌诊断的临床性能评价

目的:评估电化学发光法检测尿Cyfra21-1对膀胱癌的临床诊断价值。方法:收集2020年02月至2022年01月到深圳市罗湖区人民医院就诊的泌尿系统炎症疾病(尿路结石、尿路感染)患者尿液66例、膀胱癌患者尿液58例，收集深圳市众循精准医学研究院健康志愿者尿液37例和健康体检者尿液样本83例作为健康对照样本。参考美国临床和实验室标准协会(Clinical and Laboratory Standards Institute,CLSI)颁布的系列文件使用细胞角蛋白19片段(Cyfra21-1)测定试剂盒(电化学发光法)检测尿Cyfra21-1的最低检出限、线性范围、准确度、批内精密度、干扰试验进行性能，并初步建立正常参考区间，确定Cyfra21-1的阳性判断值，最后纳入新的外部验证队列对结果进行盲测验证。采用SPSS 25.0版本软件进行实验数据分析。结果:Cyfra21-1试剂盒的最低检出限为0.01 ng/mL；在0.1~500 ng/mL范围内，线性良好(r=0.995)；每台仪器测得的高、低浓度水平样本相对偏差均不超过±10%；高、低浓度水平的批内精密度变异系数(CV)分别为1.47%与1.28%；干扰实验表明，尿液中含有生物素(≤30 ng/mL)、总蛋白(≤10 g/dL)、HAMA(≤100 ng/mL)、吉西他滨(≤380 μg/mL)、酒石酸·长春瑞滨,98％(≤1.23 μg/mL)、厄洛替尼(≤17 μg/mL)、紫杉醇(≤67 μg/mL)、阿霉素(≤40 μg/mL)、异环磷酰胺(≤800 μg/mL)、依托泊苷(≤12 μg/mL)、碳铂(卡铂)(≤500 μg/mL)时，测定结果的相对偏差均在±10％范围内；尿液中含有类风湿因子(≤1 000 IU/mL)时，测定结果的相对回收率为108.47%；建立的正常参考区间为≤10.809 ng/mL；尿Cyfra21-1的阳性判断值为7.678 ng/mL，敏感度和特异性分别为67.20%和92.50%。盲测样本经检测，23份结果正确，正确率为63.89%。膀胱癌组与健康对照组和泌尿系统炎症疾病组检测结果比较，差异具有统计学意义(P＜0.01)；各组内性别比较，差异无统计学意义(P＞0.05)，膀胱癌组内TNM各分期比较，差异无统计学意义(P＞0.05)，膀胱癌组内肿瘤各分级比较，差异具有统计学意义(P＜0.01)，膀胱癌组内肿瘤浸润程度比较，差异具有统计学意义(P＜0.05)。结论:细胞角蛋白19片段(Cyfra21-1)测定试剂盒(电化学发光法)性能良好，符合临床检测尿Cyfra21-1要求。电化学发光法检测尿Cyfra21-1能有效区分膀胱癌和非膀胱癌样本，具备精密度高、时间短、操作自动化等优势，适合于临床应用。

Clinical performance evaluation of urine cytokeratin 19 fragment with ECLIA in bladder cancer diagnosis

Objective:To evaluate the clinical value of electrochemiluminescence (ECLIA) method for assay of urine Cyfra21-1 in the diagnosis of bladder cancer.Methods:Urine samples for testing were collected from 66 patients with urinary inflammatory diseases (urinary calculi,urinary tract infection) and from 58 patients with bladder cancer who were receiving treatment at Shenzhen Luohu People's Hospital from February 2020 to January 2022.As the control,urine samples were collected from 37 healthy volunteers and 83 healthy people indicated by a relevant physical examination from Shenzhen Following Precision Medical Research Institute.The cytokeratin 19 fragment (Cyfra21-1) assay kit (ECLIA) was observed in terms of the performance in lower limit of detection (LLOD),linear range,accuracy,within-run precision,and interference based on a series of files issued by the Clinical and Laboratory Standards Institute (CLSI).Meanwhile,the relevant normal reference interval was established,and the cutoff value for positive result of Cyfra21-1 was determined.A new external validation cohort was finally included for blind test validation of the results.Statistical analysis was performed using SPSS 25.0 version software for experimental data analysis.Results:For the Cyfra21-1 assay kit,the LLOD was 0.01 ng/mL,and in the range of 0.1~500 ng/mL,the linearity was good (r=0.995).For both high and low concentration samples measured by each instrument,the relative deviation was no higher than ±10%.The coefficients of variation (CV) of within-run precision at high and low concentrations were 1.47% and 1.28%,respectively.The results of the interference test reveal that when the urine contained biotin (≤30 ng/mL),total protein (≤10 g/dL),HAMA (≤100 ng/mL),gemcitabine (≤380 μg/mL),98% vinorelbine tartrate (≤1.23 μg/mL),erlotinib (≤17 μg/mL),paclitaxel (≤67 μg/mL),doxorubicin (≤40 μg/mL),ifosfamide (≤800 μg/mL),etoposide (≤12 μg/mL) and carboplatin (≤500 μg/mL),the relative deviations of the test results were all within ±10%,and when the urine contained rheumatoid factor (≤1 000 IU/mL),the relative recovery rate of the test results was 108.47%.The established normal reference interval was ≤10.809 ng/mL.For urine Cyfra21-1,the cutoff value for a positive result was 7.678 ng/mL,and the sensitivity and specificity were 67.20% and 92.50%,respectively.For the differences between the bladder cancer group and the healthy control group or the urological inflammatory disease group,there was statistical significance (P＜0.01).However,there was no statistical significance for the intra-group differences between males and females (P＞0.05) and the differences between TNM stages in the bladder cancer group (P＞0.05).Moreover,a statistical significance was noted in the differences between pathological grade in the bladder cancer group (P＜0.01),and the tumor infiltrating (P＜0.05).The results of the 23 samples used for blind testing were validated for correctness,and the correct rate was 63.89%.Conclusion:The Cyfra21-1 assay kit (ECLIA) has good performance,and meets the relevant requirements for clinical tests.The ECLIA for urine Cyfra21-1 can distinguish bladder cancer samples from non-bladder cancer samples effectively with the advantages of having high precision and automatic operation,as well as being time saving.Thus,said method is suitable for clinical application.