The role of misclassification in progression of CIN2 during active surveillance: results from a historical cohort

BACKGROUND: Many countries have recently implemented active surveillance for cervical intraepithelial neoplasia grade 2 (CIN2) as an option in younger women due to high regression rates and because excisional treatment is associated with increased risk of preterm birth. However, the known low reproducibility of the CIN2 diagnosis poses a major challenge for active surveillance, suggesting that the observed risk of progression in prior studies may be due to initial misclassification of the CIN2 diagnosis. OBJECTIVE: This study aimed to assess the association between expert-verified histological diagnoses of cervical biopsies in women undergoing active surveillance for CIN2 and the subsequent risk of cervical intraepithelial grade 3 or worse (CIN3+). STUDY DESIGN: We conducted a historical cohort study on women undergoing active surveillance for CIN2, diagnosed at Aarhus University Hospital from 2000-2010. Women were identified through the Danish Pathology Data Bank and were eligible for inclusion if they were 23-40 years old at the time of CIN2 diagnosis. Women with a prior history of CIN2 or worse (CIN2+), loop electrosurgical excision procedure (LEEP), or hysterectomy were excluded. Three expert pathologists independently graded cervical lesions by reviewing hematoxylin and eosin (H&E) and p16 stained tissue slides from the time of CIN2 diagnosis. Women were followed for 28 months, from the date of the CIN2 diagnosis until a record of progression, LEEP, hysterectomy, or end of follow-up, whichever occurred first. Using modified Poisson regression analysis, we estimated the crude and adjusted relative risks of CIN3+, including corresponding 95% confidence intervals (CI). We adjusted for age, index cytology, and HPV genotype as potential confounders. RESULTS: A total of 437 women were included with a median age of 27 years. Upon expert review, 56 (12.8%) were upgraded to CIN3, 261 (59.7%) had CIN2 confirmed, and 120 (27.5%) were downgraded to CIN1/normal. Overall, 173 (39.6%) women had a subsequent record of CIN3+, ranging from 22.5% in women with expert CIN1/normal, to 42.2% in women with expert CIN2, and 64.3% in women with expert CIN3. Thus, women with an expert-verified CIN3 diagnosis were more likely to have a subsequent record of CIN3+ during follow-up (adjusted relative risk, aRR 1.37 (1.07; 1.75)) compared to those who had expert CIN2. This association remained significant in stratified analyses for women aged 31-40 years (aRR 1.85 (1.25; 2.76)), in women with high-grade index cytology (aRR 1.43 (1.04; 1.97)), and women with a non-HPV16 lesion (aRR 1.66 (1.09; 2.52)). The risk of CIN3+ remained high for women presenting with a high-grade cytology or HPV16 positivity at baseline, irrespective of expert diagnosis. CONCLUSION: Our findings demonstrate substantial interobserver variability in the diagnosis of community-detected CIN2, with nearly half of the cases being reclassified upon expert review. This underscores the heterogeneity of this diagnosis and the importance of accurate diagnosis to minimize both under- and overtreatment. These results highlight the need for individualized, risk-based management strategies for CIN2 that integrate virological and cytological factors rather than relying solely on histopathology.