Factors associated with multicentric intraepithelial neoplasia of the lower genital tract: a retrospective single-center case-control study

Multicentric lower genital tract intraepithelial neoplasia involves multiple intraepithelial lesions in the vagina, cervix, and vulva and is commonly associated with persistent high-risk human papillomavirus infection. Despite improved cervical lesion screening, understanding of vulvovaginal lesions remains limited. Patients with cervical lesions often have concurrent vulvar and/or vaginal intraepithelial neoplasia, and these concurrent lesions may be overlooked without multisite assessment. To characterize the clinical features and identify factors associated with cervicovaginal-vulvar multicentric intraepithelial neoplasia and to explore the apparent performance of multivariable models in distinguishing multicentric from isolated cervical lesions. We conducted a retrospective, single-center, case-control study of 362 patients with histologically confirmed cervical lesions and concurrent vaginal and/or vulvar lesions identified during the same diagnostic episode, and 391 patients with isolated cervical lesions. Clinical and colposcopic characteristics were compared, and independent associations were assessed using multivariable logistic regression. Apparent model performance was assessed using receiver operating characteristic, calibration, and decision curve analyses. Concurrent high-grade vaginal and vulvar lesions were significantly more common among patients with high-grade than among those with low-grade cervical lesions (vaginal high-grade squamous intraepithelial lesions: 34.75% vs. 12.30%, P < 0.001; vulvar high-grade squamous intraepithelial lesions: 20.34% vs. 10.25%, P = 0.031). In multivariable analyses, an HPV DNA signal of ≥ 500 RLU/CO, HPV16/18-group positivity, absence of partner circumcision, and absence of HPV vaccination were independently associated with low-grade multicentric involvement. Factors independently associated with high-grade multicentric involvement included menopause, sexual debut ≤ 18 years, an HPV DNA signal of ≥ 500 RLU/CO, persistent HPV infection of ≥ 1 year, absence of partner circumcision, absence of HPV vaccination, and HPV16/18 positivity. The exploratory low-grade model demonstrated modest apparent discrimination (area under the receiver operating characteristic curve = 0.703), whereas the high-grade model showed moderate apparent discrimination (area under the receiver operating characteristic curve = 0.777) in the development dataset. High-grade cervical lesions were associated with more frequent concurrent high-grade vaginal and vulvar lesions, highlighting the importance of comprehensive multisite assessment. Several demographic, behavioral, and HPV-related characteristics were associated with multicentric involvement, whereas HPV vaccination and partner circumcision showed inverse associations. Because the study was retrospective and the models were not independently validated, these findings should be considered hypothesis-generating and may support closer multisite assessment rather than causal inference or immediate clinical prediction. Multicentric involvement was associated with HPV16/18-group positivity, a higher HPV DNA signal, the absence of HPV vaccination, and the absence of partner circumcision. The exploratory models displayed a modest-to-moderate apparent discrimination and require internal and external validation.