[Performance of point-of-care testing for cervical cancer screening].

Objective: To evaluate the performance of point-of-care testing for cervical cancer and precancerous lesions screening. Methods: In September 2020, 197 and 273 women were selected by using simple random sampling method from "self-sampling" cohort and "physician-sampling" cohort established in Xiangyuan county, Shanxi Province, China, respectively. Cervical exfoliated cells were collected by women themselves or gynecologists. All samples were detected by POCT and women with positive result were directly referred for colposcopy. Subsequently, all the samples were detected by careHPV and PCR test. Colposcopy and punch biopsy were performed for women with POCT negative but careHPV or PCR test positive at another visit. Using histopathological diagnosis as the gold standard, we calculated sensitivity, specificity and drew the receiver operating characteristic (ROC) curves. The accuracy of POCT was analyzed and compared to that of careHPV and conventional PCR test in cervical cancer and precancerous lesions screening. Results: The median (Q1, Q3) age of 470 women was 51 (45, 57) years old. Based on self-sampling, the sensitivity and specificity of POCT for CIN2+ were 100.00% (95%CI: 56.56%-100.00%) and 28.95% (95%CI: 22.97%-35.76%), respectively. Compared with POCT, POCT HPV16/18 test had similar sensitivity and higher specificity of 89.47% (95%CI: 84.30%-93.08%). Self-sampling POCT HPV16/18 test had an AUC of 0.947 (95%CI:0.910-0.985), which was higher than that of careHPV and PCR test. Physician-sampling POCT test had 100.00% sensitivity (95%CI: 64.57%-100.00%) and 55.85% specificity (95%CI: 49.83%-61.70%) for detecting CIN2+. POCT HPV16/18 test had lower sensitivity (71.43%, 95%CI: 35.90%-91.76%) and higher specificity (92.45%, 95%CI: 88.63%-95.06%). POCT HPV16/18 test generally showed similar AUC on both self-collected samples and clinician-collected samples (0.947 vs 0.819, P=0.217). Conclusion: POCT HPV16/18 test is an effective method with relatively high sensitivity and specificity for cervical cancer screening.目的： 分析即时检测（POCT）技术筛查宫颈癌及癌前病变的准确性。 方法： 2020年9月，在山西省襄垣县建立的“自我采样”队列和“医生采样”队列中，采用简单随机抽样的方法从两队列人群中抽取研究对象，分别纳入197和273名。由研究对象自己或妇科医生分别采集宫颈脱落细胞标本并进行检测，采用POCT检测人乳头状瘤病毒（HPV），检测阳性者直接行阴道镜检查；同时采用careHPV及常规PCR检测HPV，对POCT检测阴性但careHPV或常规PCR检测任一阳性者再次进行阴道镜检查和活检。以组织病理学诊断为金标准，绘制受试者工作特征（ROC）曲线，计算灵敏度、特异度，比较POCT与careHPV及常规PCR筛查宫颈癌及癌前病变的准确性。 结果： 470名对象的年龄M（Q1，Q3）为51（45，57）岁。将组织病理学诊断宫颈上皮内瘤变2级及以上病变（CIN2+）作为疾病终点，基于自我采样方式的POCT检测的灵敏度及特异度分别为100.00%（95%CI：56.56%~100.00%）、28.95%（95%CI：22.97%~35.76%），POCT HPV16/18检测的灵敏度与POCT检测相同，特异度增加至89.47%（95%CI：84.30%~93.08%）；与careHPV和常规PCR相比，自我采样POCT HPV16/18检测的AUC最佳，为0.947（95%CI：0.910~0.985）。基于医生采样方式的POCT检测的灵敏度及特异度分别为100.00%（95%CI：64.57%~100.00%），55.85%（95%CI：49.83%~61.70%）；POCT HPV16/18检测的灵敏度降低至71.43%（95%CI：35.90%~91.76%），特异度增加至92.45%（95%CI：88.63%~95.06%）。基于自我采样方式的POCT HPV16/18检测的AUC和医生采样方式差异没有统计学意义（0.947比0.819，P=0.217）。 结论： POCT HPV16/18检测灵敏度和特异度较高，是基于人群的宫颈癌筛查可选择的有效方法。.