Analysis of factors influencing the missed diagnosis of cervical precancerous lesions by colposcopy

OBJECTIVE: We aimed to explore the factors that influence the missed diagnosis of cervical precancerous lesions by colposcopy. METHODS: This retrospective cohort study included 448 patients with cervical precancerous lesions who underwent colposcopy and cervical biopsy between January 1, 2022, and October 13, 2024. Univariate and multivariate logistic regression analyses were performed to identify factors associated with missed diagnosis, with colposcopic findings compared against histopathological results as the gold standard. RESULTS: Compared to histopathological results (gold standard), the accuracy rate of colposcopy was 65.63% (294/448). Univariate logistic regression analysis showed that insufficient colposcopic diagnosis was significantly correlated with age ≥40 years (χ2=15.632, p<0.001), absence of postcoital vaginal bleeding (χ²=5.734, p=0.01), low-grade abnormalities in liquid-based cytology (χ²=9.809, p=0.002), non-HPV16 infection (χ2=7.204, p=0.007), lesions located in the cervical canal (χ²=23.914, p<0.001), history of cervical conization (χ2=7.182, p=0.007), and colposcopic features including no or thin acetowhite epithelium (χ²=225.559, p<0.001), absence of mosaic (χ2=45.925, p<0.001), absence of punctate blood vessels (χ²=36.041, p<0.001), and positive iodine test (χ2=35.794, p<0.001). Multivariate logistic regression analysis identified age ≥40 years (OR=0.468, 95% CI: 0.287~0.761, p=0.002), absence of postcoital vaginal bleeding (OR=2.132, 95% CI: 1.077~4.222, p=0.030), low-grade abnormalities in TCT results (OR=2.584, 95% CI: 1.527~4.372, p<0.001), non-HPV16 infection (OR=0.545, 95% CI: 0.358~0.832, p=0.005), and lesions localized to the cervical canal (OR=0.042, 95% CI: 0.005~0.338, p=0.003) as independent risk factors for missed diagnosis. The highest missed diagnosis rate (67.9%) was observed in patients aged ≥55 years with low-grade TCT abnormalities (n=53), which should be interpreted with caution as exploratory data; in contrast, the missed diagnosis rate was 20.0% in patients aged <25 years with HPV16 positivity (n=15). CONCLUSION: Age over 40, absence of postcoital vaginal bleeding, low-grade abnormality in TCT results, non-HPV16 infection, and lesions in the cervical canal are independent risk factors for cervical precancerous lesions detected through colposcopy. However, the findings must be interpreted cautiously because they were based on a retrospective study and the study had a short follow-up.