Evaluation of CIN2/3 Lesion Regression in GynTect® DNA Methylation-Marker-Negative Patients in a Longitudinal Study

Cervical intraepithelial neoplasia (CIN) grade 2/3 has a high spontaneous regression rate, especially among women ≤29 years of age. To reduce overtreatment, reliable prognostic biomarkers would be helpful. The main aim of this study was to analyze the negative predictive value of the methylation marker panel GynTect® for lesion regression. In this prospective, multicenter, longitudinal observational proof-of-concept study, women aged ≤29 years with histologically confirmed CIN2 (n = 24) or CIN3 (n = 36) were closely monitored without treatment for up to 24 or 12 months, respectively. The outcome was either regression, persistence, or progression of the lesion. For each patient, a single baseline sample (V0) for cytology, hrHPV detection and methylation analysis was taken. In a primary analysis, the negative predictive value (NPV) of a GynTect®-negative test result at V0 for regression was determined. We tested the null hypothesis NPV ≤ 70% against the alternative hypothesis NPV ≥ 90%. Twelve of the eighteen GynTect®-negative CIN2 patients showed regression (NPV = 67%, 90% CI 44-85%, p = 0.53). Of the 27 GynTect®-negative CIN3 lesions, 15 regressed (NPV = 56%, 90% CI 38-72%, p = 0.92). Although the majority of GynTect®-negative lesions regressed, the postulated NPV of ≥90% was not observed. Thus, the clinical relevance for an implementation of the GynTect® assay for patients undergoing watchful waiting remains questionable. Further studies with longer observation periods should be undertaken.

Evaluating the Feasibility of Machine-Learning-Based Predictive Models for Precancerous Cervical Lesions in Patients Referred for Colposcopy

Background: Colposcopy plays an essential role in cervical cancer control, but its performance remains unsatisfactory. This study evaluates the feasibility of machine learning (ML) models for predicting high-grade squamous intraepithelial lesions or worse (HSIL+) in patients referred for colposcopy by combining colposcopic findings with demographic and screening results. Methods: In total, 7485 patients who underwent colposcopy examination in seven hospitals in mainland China were used to train, internally validate, and externally validate six commonly used ML models, including logistic regression, decision tree, naïve bayes, support vector machine, random forest, and extreme gradient boosting. Nine variables, including age, gravidity, parity, menopause status, cytological results, high-risk human papillomavirus (HR-HPV) infection type, HR-HPV multi-infection, transformation zone (TZ) type, and colposcopic impression, were used for model construction. Results: Colposcopic impression, HR-HPV results, and cytology results were the top three variables that determined model performance among all included variables. In the internal validation set, six ML models that integrated demographics, screening results, and colposcopic impression showed significant improvements in the area under the curve (AUC) (0.067 to 0.099) and sensitivity (11.55% to 14.88%) compared with colposcopists. Greater increases in AUC (0.087 to 0.119) and sensitivity (17.17% to 22.08%) were observed in the six models with the external validation set. Conclusions: By incorporating demographics, screening results, and colposcopic impressions, ML improved the AUC and sensitivity for detecting HSIL+ in patients referred for colposcopy. Such models could transform the subjective experience into objective judgments to help clinicians make decisions at the time of colposcopy examinations.

Population-based trends in high-grade cervical lesions in the early human papillomavirus vaccine era in the United States

BACKGROUND

Cervical intraepithelial neoplasia grade 2, 3, and adenocarcinoma in situ (CIN2+) lesions can be monitored as early indicators of human papillomavirus (HPV) vaccine impact. Changes to screening utilization will affect observed reductions in CIN2+ rates and complicate the interpretation of vaccine impact.

METHODS

From 2008 to 2012, 9119 cases of CIN2+ among 18- to 39-year-old residents of catchment areas in California, Connecticut, New York, and Oregon were reported to the HPV-IMPACT Project, a sentinel system for monitoring the population impact of HPV vaccine. Age-stratified CIN2+ incidence rates were calculated for each catchment. Annual cervical screening was estimated for California, New York, and Oregon catchments with administrative and survey data. The Cochran-Armitage test was used to examine trends.

RESULTS

From 2008 to 2012, the incidence of CIN2+ significantly decreased among 18- to 20-year-olds (California, from 94 to 5 per 100,000 women; Connecticut, from 450 to 57 per 100,000 women; New York, from 299 to 43 per 100,000 women; and Oregon, from 202 to 37 per 100,000 women; Ptrend  < .0001) and among 21- to 29-year-olds in Connecticut (from 762 to 589 per 100,000 women) and New York (from 770 to 465 per 100,000 women; Ptrend  < .001); rates did not differ among 30- to 39-year-olds. During the same period, screening rates also declined, with the largest decreases among 18- to 20-year-olds (from 67% in Oregon to 88% in California) and with smaller declines among 21- to 29-year-olds (13%-27%) and 30- to 39-year-olds (3%-21%).

CONCLUSIONS

The declines in CIN2+ detection in young women were likely due to reduced screening but could also reflect the impact of vaccination. These data illustrate challenges in interpreting CIN2+ ecologic trends in the new era of cervical cancer prevention and emphasize the importance of information such as HPV types detected in lesions to assess the impact of HPV vaccine on cervical precancers.