GY66阴道微生态形态学全自动检测仪临床应用评价

翟青枝; 李立安; 王雪琪; 范一帆; 于东渤; 付玉荣; 李江华; 姜淑芳

doi:10.3969/j.issn.2095-5227.2022.04.012

GY66阴道微生态形态学全自动检测仪临床应用评价

Clinical evaluation of GY66 automatic vaginal microecological morphology detector

摘要

摘要:
背景阴道微生物形态学检测是阴道感染性疾病的主要检测方法，以人工镜检为主，耗时较长且主观偏倚明显。基于深度学习的全自动阴道微生态检测仪可有效解决现有问题，但投入使用前需要严谨的对照试验进行验证。
目的分析GY66形态学全自动检测仪在阴道微生态检测中的有效性。
方法收集2020年12月1日- 2021年7月31日解放军总医院第一医学中心妇产科门诊患者的阴道分泌物样本303例，分别采用全自动GY66阴道微生态检测仪与革兰染色人工镜检进行滴虫、霉菌、线索细胞及清洁度比较。以人工镜检结果为金标准，评价全自动GY66检测仪阴道微生态有形成分检测的性能。
结果全自动GY66检测仪与人工镜检的霉菌、线索细胞、滴虫总体检出率差异无统计学意义(P＞0.05)。全自动GY66检测仪的霉菌、滴虫、线索细胞和清洁度检出的总符合率分别为98.68%、99.67%、99.67%和97.69%。全自动GY66检测仪与人工镜检霉菌、线索细胞、滴虫和清洁度的检测结果一致性程度强(Kappa值分别为0.961、0.981、0.939和0.950)，差异无统计学意义。抗干扰能力及交叉污染试验中高浓度阳性样本均不会影响下一个空白标本的结果。抽检3份样本各做3次GY66仪器镜检，重复性检测结果一致。
结论 GY66全自动阴道微生态检测仪进行阴道分泌物检查具有良好的有效性，可为临床提供可靠准确的镜检结果。

Abstract:
Background Vaginal microbial morphology detection is the main detection method for vaginal infectious diseases, mainly by artificial microscopy, which is time-consuming and has obvious subjective bias. The existing problems can be solved effectively by the automatic vaginal microbial morphology detector based on deep learning, but rigorous controlled test verification is necessary before it is put into use.
Objective To analyze the effectiveness of GY66 automatic detector in vaginal microecological morphological detection.
Methods Totally 303 samples of vaginal secretions were collected from outpatients in the Department of Obstetrics and Gynecology of the First Medical Center of Chinese PLA General Hospital from December 1, 2020 to July 31, 2021. The samples were detected by two detection methods, that were automatic GY66 vaginal microecological detector and manual microscopy after Gram stain. And the detection results were compared, included trichomonas, mold, clue cells and cleanliness. Taking the manual microscopy results as gold standard, the clinical performance of automatic GY66 detector for detecting vaginal microecological formed components was evaluated.
Results There was no statistic difference in the overall detection rates of mold, clue cells and trichomonas between automatic GY66 detector and manual microscopy (P＞0.05). The total coincidence rates of mold, trichomonas, clue cells or cleanliness detected by automatic GY66 detector were 98.68%, 99.67%, 99.67% and 97.69% respectively, which were highly consistent with manual microscopy (Kappa values were 0.961, 0.982, 0.939 and 0.950) with no statistical significance. High-concentration positive samples in anti-jamming and cross-contamination tests did not affect the results of the next blank sample. The microscopic examination by GY66 detector was performed in 3 random samples, 3 times for each sample, and the repeatability test results were consistent.
Conclusion The automatic vaginal microecological detection method based on deep learning has good effectiveness in vaginal secretion examination, which can quickly provide reliable and accurate microscopic examination results for clinical practice.

HTML全文

参考文献(19)

施引文献

资源附件(0)