[Analysis of factors influencing the false-negative diagnosis of cervical/vaginal liquid based cytology]

Objective: To investigate the possible influencing factors of false-negative diagnosis of cervical/vaginal liquid based cytology, and further improve the sensitivity of cervical/vaginal cytology. Methods: The results of cervical/vaginal cytology of outpatients and inpatients in Department of Obstetrics and Gynecology, Peking University First Hospital from July 2015 to December 2018 were analyzed retrospectively. Cytological false-negative cases were defined as the patients whose cytological results showed no intraepithelial neoplasia and malignant (NILM), but whose biopsy was diagnosed as cervical intraepithelial neoplasia (CIN) 2 or vaginal intraepithelial neoplasia (VAIN) 2 or above within 6 months of the diagnosis. The review of false-negative cytology smear was completed by two senior cytologists. Two-class logistic regression was used to evaluate the influence of age, location or number of lesion, and degree of lesion on the false-negative diagnosis of cytology. The reasons for the inconsistency of false-negative diagnosis were analyzed with the review results. Results: Among 1 009 cases of CIN2+ and VAIN2+ lesions, 180 cases (17.8%) showed NILM. After reviewing the smear, 123 cases (68.3%) were identified as NILM and 57 cases(31.7%) as abnormal. The false-negative rate was the highest (20.8%) in the patients with age≤30 years, and the risk was 8.85 times higher than the patients aged 31 to 60 years (P<0.001), 9.26 times than the patients aged≥60 years (P<0.001). The highest cytological false-negative rate was 50.0% for cervical polyps or intraductal lesions. The false-negative rate of vaginal wall or vaginal pedicle rupture was 13.0%; that of single cervical lesion was 22.3%; that of high-grade squamous intraepithelial lesion(HSIL) and adenocarcinoma in situ of cervix(AIS) was 13.7% and that of malignant lesions was 3.9%. The most common cell types in the reviewed abnormal cases were squamous cells in the middle surface layer (38.6%) and squamous cells in the outer bottom layer (24.6%). The abnormal cells in all smears was the most common distribution (59.7%), the number of abnormal cells in the smear was less than 10 (31.6%), nuclear enlargement and light staining were common (42.2%), and inflammatory lesions or keratotic changes in the background were most common (59.7%). Conclusions: Age of the patient, location or number of lesion, and degree of lesion are associated with false-negative diagnosis of cytology. Summarizing sampling experience and improving sampling skills will help reduce the occurrence of false-negative cases. Cytopathologists should examine slightly abnormal changes more carefully and learn how to further reduce the false-negative rate procactively.