LEEP, also known as large loop excision of the transformation zone, is one of the most widely used treatments for high-grade cervical lesions worldwide. The procedure removes the abnormal transformation zone with an electrified wire loop, allowing both treatment and histopathologic assessment in a single intervention. Surgeons have long faced a difficult trade-off: cutting deeper may improve the chances of removing all diseased tissue, but more extensive excision is associated with an increased risk of adverse obstetric outcomes, including preterm birth, a concern that is particularly relevant for women of reproductive age.

To investigate whether excision depth carries independent prognostic information, the team analyzed deidentified clinical and pathology records from women who underwent LEEP for newly diagnosed CIN2/3 at their institution between April 2018 and November 2025. Of 1,363 unique patient candidates identified, 488 had complete data for the core exposure and covariates. The primary margin analysis included 422 women, of whom 41, or 9.7 percent, had positive excision margins. The postoperative hrHPV analysis included 275 women, of whom 61, or 22.2 percent, had a recorded positive result after surgery.

The researchers evaluated excision depth as a continuous variable, reporting effect estimates per 3-millimeter increase rather than using arbitrary cutoffs. In the core multivariable model, adjusted for age, histopathologic grade, and preoperative HPV16/18 status, each additional 3 millimeters of depth was associated with an odds ratio of 0.95 for margin positivity (95 percent confidence interval, 0.72 to 1.26; P = 0.74), showing no statistically meaningful relationship. For recorded postoperative hrHPV positivity, the adjusted odds ratio per 3-millimeter increase was 0.80 (95 percent confidence interval, 0.57 to 1.13; P = 0.20), again failing to reach significance.

By contrast, the histologic grade of the lesion showed a strong and consistent association with margin status. Women with CIN3 had roughly 3.7-fold higher odds of positive margins compared with those with CIN2 (adjusted odds ratio, 3.74; 95 percent confidence interval, 1.90 to 7.34; P < 0.001). Increasing age was also linked to higher odds of margin positivity, though more modestly, with an adjusted odds ratio of 1.35 per 10-year increase (95 percent confidence interval, 1.00 to 1.82; P = 0.047). For postoperative hrHPV positivity, age was the standout factor: each additional decade of age raised the odds by 66 percent (adjusted odds ratio, 1.66; 95 percent confidence interval, 1.24 to 2.21; P < 0.001), a pattern the authors note is consistent with prior work implicating age-related transformation-zone retraction and immunosenescence in post-treatment viral persistence.

The robustness of these null findings was tested through an unusually thorough battery of sensitivity analyses. Natural cubic spline modeling found no departure from linearity in either outcome. Inverse-probability-of-observation weighting, designed to correct for the substantial loss of follow-up data, produced estimates nearly identical to the complete-case results. A specification curve analysis spanning 27 different model combinations, varying margin definitions, depth caps, and adjustment strategies, yielded a median odds ratio of 0.95 with individual estimates ranging from 0.82 to 1.09, and none of the 27 specifications met the false-discovery-rate-adjusted threshold for significance. Interaction tests across ten prespecified clinical subgroups likewise found no evidence of effect modification.

An exploratory pathway analysis among 236 patients with both margin and hrHPV data compared hypothetical excision depths of 8 millimeters versus 5 millimeters. The standardized difference in predicted postoperative hrHPV positivity was −5.80 percentage points in total, but when margin status was accounted for, the direct difference remained essentially unchanged at −5.71 percentage points, while the margin-mediated indirect component was nearly zero at −0.08 percentage points. The authors interpret this as suggesting that any relationship between depth and postoperative viral detection is unlikely to operate through margin status, though they emphasize this analysis is hypothesis-generating rather than evidence of a causal pathway.

The study’s context matters for interpreting the results. The median excision depth in the cohort was just 6 millimeters, with an interquartile range of 5 to 8 millimeters and a full observed range of 2 to 35 millimeters. Relatively few patients underwent very deep resection, so the null findings apply strictly to routine, standard-practice LEEP cases within this narrow depth range. The authors explicitly caution that their results cannot be extrapolated to clinical scenarios requiring deep endocervical excision, such as lesions associated with type 3 transformation zones, for which deeper conization remains clinically necessary and indicated.

Several limitations temper the conclusions. Postoperative hrHPV testing was not standardized, and the intervals between surgery and testing varied widely, so recorded postoperative positivity serves only as an exploratory marker of viral detection rather than verified persistent infection; new incident infections cannot be excluded. Genotype-specific tracking between pre- and postoperative samples was not possible because archived records did not consistently store full genotype outputs. Sample attrition was substantial, dropping from 1,363 candidates to 488 in the core cohort, and although weighting analyses addressed missingness related to measured covariates, residual selection bias and confounding by indication cannot be ruled out. Variables such as transformation-zone type, lesion dimensions, parity, smoking status, loop size, and operator characteristics were unavailable, and pathology-recorded specimen depth may not exactly reflect the intended in vivo depth due to tissue shrinkage and handling.

Taken together, the findings argue against relying on excision depth alone as a marker of treatment adequacy in routine LEEP for CIN2/3. The authors frame their results as supporting a broader clinical assessment that integrates lesion characteristics, transformation-zone anatomy, reproductive considerations, margin interpretation, and guideline-based postoperative HPV surveillance, rather than as a recommendation to alter current surgical practice. Future prospective studies with standardized HPV testing intervals, detailed anatomical characterization, and longer-term recurrence outcomes would help clarify whether excision depth holds greater relevance in specific patient or anatomical subgroups.

Subject of Research: Association between LEEP excision depth and margin status or postoperative hrHPV positivity in women treated for CIN2/3

Article Title: LEEP excision depth is not independently associated with margin positivity or recorded postoperative hrHPV positivity in women with CIN2/3: a retrospective cohort study

Article References: Cheng, J., Hu, C., Zhu, G., Liu, F., & Li, G. (2026). LEEP excision depth is not independently associated with margin positivity or recorded postoperative hrHPV positivity in women with CIN2/3: a retrospective cohort study. Frontiers in Oncology, 16, 1969282. https://doi.org/10.3389/fonc.2026.1969282

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1969282

Keywords: LEEP, CIN2/3, cervical intraepithelial neoplasia, excision depth, margin status, high-risk HPV, HPV persistence, retrospective cohort, cervical cancer prevention, loop electrosurgical excision, gynecologic oncology, preterm birth risk