Is active surveillance of CIN2 always appropriate? Clinical outcomes and factors associated with progression and regression over 48 months in an HPV-positive cohort Authors: Maria Teresa Bruno, Alessia Pagana, Tiziana Aiello, Elena Pivetti, Antonino Giovanni Cavallaro, Mattia Tarascio, Marco Marzio Panella, Gaetano Valenti Published: 2026 ABSTRACT AbstractBackgroundActive surveillance of cervical intraepithelial neoplasia grade 2 (CIN2) is increasingly adopted to reduce the risk of overtreatment, particularly in young women. However, the appropriateness of this approach remains debated, especially in light of evidence suggesting an increased long-term risk of cervical cancer in certain subpopulations.MethodsA retrospective cohort study was conducted on HPV-positive women with histologically confirmed p16-positive CIN2 managed with active surveillance. Medium-term clinical outcomes included regression, persistence, and progression to CIN3 or higher during follow-up. Survival analyses (Kaplan–Meier), Cox regression models, and Modified Poisson regression models with robust standard errors were performed to identify independent factors associated with progression. The mean follow-up was 27 months, with a maximum follow-up of 48 months.ResultsDuring the observation period, spontaneous regression was observed in 61% of cases, while 25.5% of patients showed progression and 13.8% persistence of the lesion. Regression occurred predominantly within the first 24–36 months, followed by a plateau in the curves. Progression was significantly associated with high-grade cytology (ASC-H/HSIL) and HPV16/18 positivity. In multivariable models, both factors remained independent factors associated with progression, whereas non-16/18 HPV genotypes and low-grade cytology were associated with a higher likelihood of regression.ConclusionsAlthough this study does not allow for a direct assessment of long-term cancer risk, the identification of early factors associated with persistence and progression provides clinically relevant insights into CIN2 profiles associated with an increased risk of progression within the observed follow-up period. These findings suggest that active surveillance of CIN2 may not be equally appropriate across all patient subgroups and requires careful risk stratification based on virological and cytological factors to support selective and clinically sound management.Keywords: CIN2, Active surveillance, HPV16, Regression, Progression, Risk stratification, Cervical cancer risk Introduction Active surveillance of cervical intraepithelial neoplasia grade 2 (CIN2) has progressively emerged in recent years as a management strategy alternative to immediate excisional treatment, particularly in young and reproductive-aged women [1–6]. This approach is based on evidence that a significant proportion of CIN2 lesions may undergo spontaneous regression, thereby reducing the risk of overtreatment and the potential obstetric complications associated with procedures such as LEEP or conization [7–9]. However, the increasing adoption of active surveillance has raised a central question: is this strategy equally appropriate for all patients? While short- to medium-term regression rates are encouraging, population-based studies with very long follow-up have suggested an increased cumulative risk of invasive cervical cancer in women with conservatively managed CIN2 compared with those undergoing excisional treatment [10]. These observations suggest that the biological behavior of CIN2 is not uniform and that histological diagnosis alone is not sufficient to define the true evolutionary potential of the lesion [11]. Cervical cancer represents the end result of a multifactorial and temporally prolonged process, primarily driven by the persistence of high-risk HPV infection [12–14]. Within this biological continuum, CIN2 occupies an intermediate and controversial position: it may represent a lesion undergoing spontaneous regression or a dynamic phase of progression toward higher-grade lesions [15]. The distinction between these two trajectories is not always evident at the time of initial diagnosis. It is well documented that persistence of HPV16/18 infection is associated with an increased long-term risk of CIN3+, with cumulative rates of up to 40% at 10 years [16, 17]. Similarly, the presence of HSIL cytology in association with CIN2 has been correlated with a higher likelihood of an underlying higher-grade lesion and a greater probability of progression [18, 19]. These elements indicate that CIN2 should not be considered a static entity, but rather a biologically heterogeneous condition whose clinical significance depends on the patient’s virological and cytological profile [11]. In light of this evidence, the assessment of the safety of active surveillance cannot be limited to short-term regression rates alone but must necessarily consider the potential future oncologic impact. In this context, the early identification of factors associated with persistence and progression assumes crucial importance, as these factors may represent surrogate markers of increased risk over time. Our study therefore aims to achieve a dual objective: on the one hand, to evaluate the clinical outcomes of CIN2 in an HPV-positive cohort managed with active surveillance up to 48 months; on the other hand, to identify the clinical and virological profiles associated with a higher probability of progression, in order to better define subgroups with different risk profiles over time. In this way, this work aims to contribute to a risk-based management of CIN2, moving beyond the binary “treat versus observe” approach and supporting a more rigorous selection of patients eligible for active surveillance. Materials and methods A multicenter retrospective observational study was conducted on HPV-positive women with a histological diagnosis of cervical intraepithelial neoplasia grade 2 (CIN2) confirmed by p16-positive immunohistochemistry, managed with an active surveillance protocol at the University of Catania. All histological specimens were independently reviewed by two expert gynecologic pathologists, blinded to clinical data. Only cases with concordant diagnoses of p16-positive CIN2 were included, in accordance with international recommendations indicating p16 as a confirmatory diagnostic criterion. Women aged between 25 and 40 years, positive for high-risk HPV (including HPV 16/18 and other HR-HPV genotypes), with abnormal baseline cytology and a fully visible transformation zone (TZ1 or TZ2), with complete visualization of the squamocolumnar junction and lesion margins, were considered eligible. A minimum follow-up of 12 months or until the achievement of a defined clinical endpoint was required. Patients with glandular lesions (adenocarcinoma in situ or glandular dysplasia), previous excisional cervical treatment, or incomplete follow-up data were excluded. The final study population consisted of 274 women, with a mean age of 32.4 ± 5.1 years. The active surveillance protocol included colposcopic and cytological assessments every six months and annual HPV testing. In case of cytological or colposcopic worsening, a targeted biopsy was performed. A control biopsy was also systematically performed at 24 months in order to reassess the histological status of the lesion in cases of persistence. This evaluation was intended to exclude progression to CIN3 + and to support clinical decision-making. In the absence of progression, continuation of active surveillance beyond 24 months was considered in selected patients, taking into account clinical, virological, and cytological findings, as well as patient counseling. In the presence of histological progression to CIN3 or adenocarcinoma in situ, or in cases of persistent worsening, excisional treatment was indicated. The maximum duration of surveillance was 48 months. The primary outcome of the study was histological progression to CIN3 or higher (CIN3+). Secondary outcomes included lesion regression (defined as histological regression to ≤CIN1 with two consecutive negative co-tests (cytology and HPV testing) performed at 6-month intervals) and persistence of CIN2 beyond 24 months. A confirmatory biopsy was not systematically required in all cases of apparent regression, but was performed in cases with discordant findings or clinical suspicion. Time to regression and time to progression were calculated from the date of the initial CIN2 diagnosis to the first documented event. HPV genotyping was performed on ectocervical and endocervical samples collected in ThinPrep medium, using the INNO-LiPA HPV Genotyping Extra II test (Fujirebio Inc., Japan), which allows the identification of 28 viral genotypes. For analytical purposes, patients were divided into two groups: HPV 16/18 and non-16/18 HPV, given the well-established higher oncogenicity of the former. This classification was based on the well- established higher oncogenic potential of HPV16 and HPV18 compared to other high-risk genotypes. Separate analysis of individual non-16/18 high-risk HPV types (e.g., HPV31, 33, 45) was not performed due to sample size limitations and to preserve statistical robustness. Statistical analysis was conducted using R software (version 4.3.1). Time-to-event analyses were performed using Kaplan–Meier survival curves, with group comparisons assessed by the log-rank test. Because lesion status was assessed at scheduled follow-up visits rather than continuously, event times should be interpreted within the context of interval-based follow-up assessments. Multivariable Cox regression models were applied to evaluate associations between clinical variables and regression outcomes, with results expressed as hazard ratios (HR) and 95% confidence intervals. Modified Poisson regression models with robust standard errors were used to evaluate associations between baseline clinical variables and regression or progression outcomes, with results expressed as risk ratios (RR) and 95% confidence intervals. All statistical tests were two-sided, and a p-value < 0.05 was considered statistically significant. The study was conducted in accordance with the Declaration of Helsinki (2013). The protocol was submitted to the Catania 1 Ethics Committee of the University Hospital of Catania, which approved the study and deemed written informed consent unnecessary, as this was a retrospective analysis of anonymized clinical data. Results The final cohort consisted of 274 HPV-positive women with a histologically confirmed diagnosis of p16-positive CIN2, followed under an active surveillance protocol for a mean duration of 27 months, with a maximum follow-up of 48 months. During the observation period, 168 patients (61%) showed spontaneous regression of the lesion, 36 (13.8%) had histological persistence of CIN2, and 70 women (25.5%) progressed to CIN3. All patients who progressed to CIN3 (n = 70) underwent excisional treatment (LEEP) shortly after histological diagnosis. No cases of invasive carcinoma were observed during the follow-up period. (Table 1) Table 1Clinical outcomes of the cohort (maximum follow-up: 48 months) Clinical OutcomesN° patientsPercentage (%)Regression16861%Persistence3613.8%Progression7025.5%Total274100% Clinical outcomes of the cohort (maximum follow-up: 48 months) Stratified analysis by clinical characteristics revealed substantial differences in the evolutionary behavior of the lesions. Progression was significantly more frequent among patients with baseline ASC-H/HSIL cytology compared to those with ASCUS/LSIL (32.6% vs. 19.0%). Similarly, women positive for HPV16/18 showed a higher progression rate compared to patients with other high-risk genotypes (32.5% vs. 19.6%). (Table 2) Table 2Distribution of outcomes by clinical variablesN° womenRegressionPersistenceProgressionAge 25–3011272 (64.3%)14 (12.5%)26 (23.2%) 31–4016296 (59.3%)22 (13.6%)44 (27.2%)Cytology ASCUS/LSIL142100 (70.4%)15 (10.6%)27 (19.0%) ASC-H / HSIL13268 (51.5%)21 (15.9%)43 (32.6%)HPV Genotype HPV 16/1812665 (51.6%)20 (15.9%)41 (32.5%) HPV no16/18148103 (69.6%)16 (10.8%)29 (19.6%)Lesion (quadrants) 1–2 quadrants171101 (59.1%)28 (16.4%)42 (24.6%) 3–4 quadrants10367 (65.0%)8 (7.8%)28 (27.2%)Outcomes are presented separately for each clinical variable. Within each variable, subgroup categories are mutually exclusive. However, counts reported for different variables (e.g., cytology, HPV genotype, and lesion extent) are not directly additive and should not be summed across variables Distribution of outcomes by clinical variables Outcomes are presented separately for each clinical variable. Within each variable, subgroup categories are mutually exclusive. However, counts reported for different variables (e.g., cytology, HPV genotype, and lesion extent) are not directly additive and should not be summed across variables Initial colposcopic extent of the lesion showed similar progression rates in lesions involving 3–4 quadrants compared with more limited lesions (27.2% vs. 24.6%). Consistently, lesion extent was not significantly associated with progression in multivariable analyses. (Table 2) Time-to-event analyses using Kaplan–Meier curves allowed us to describe the temporal patterns of regression and progression. Regression occurred predominantly in the first 24–36 months of follow-up, with a marked reduction in the cumulative probability of further regression beyond this interval, resulting in a plateau. Conversely, progression showed a more linear distribution over time, with events documented even beyond 36 months. (Fig. 1) Kaplan–Meier curve showing time to regression of CIN2 during active surveillance. The number-at-risk table below the x-axis reports the number of patients remaining under observation at each scheduled follow-up time point (0, 12, 24, 36, and 48 months) Histological progression to CIN3 was observed in 70 patients (25.5%) during follow-up. Progression was assessed as a clinical endpoint rather than through time-to-event analysis. Stratified analysis by viral genotype showed a significant difference in regression times: non-16/18 HPV- positive patients showed a faster and higher cumulative probability of regression than 16/18 HPV-positive patients (log-rank p = 0.0001). (Fig. 2) Kaplan–Meier curves for CIN2 regression according to HPV genotype. The curves compare the cumulative probability of CIN2 regression between women with HPV16/18 infection and those with other high-risk HPV genotypes (non-16/18) during follow-up Similarly, low-grade cytology was associated with earlier regression than high-grade cytology (log-rank p < 0.001). (Fig. 3) Kaplan–Meier curves for CIN2 regression according to baseline cytology. The curves illustrate the cumulative probability of CIN2 regression stratified by baseline cytological findings (ASC-H/HSIL vs. ASCUS/LSIL) In the multivariate Cox model, low-grade cytology and absence of HPV16/18 were confirmed to be associated with a higher probability of regression (HR 1.93; 95% CI 1.42–2.64; p < 0.001 and HR 1.94; 95% CI 1.42–2.66; p < 0.001, respectively). The reported C-index should be interpreted cautiously as an exploratory descriptive measure rather than evidence of validated predictive performance. (Table 3). Table 3Multivariate cox modelVariableHR (Hazard Ratio)95% CIp-valueCytology(ASCUS/LSIL) vs. ASC-H/HSIL1.931.42–2.64< 0.001HPV(non-16/18) vs. 16/181.941.42–2.66< 0.001Concordance (C-index): 0.683. Likelihood ratio test: χ² = 38.99, p < 0.001 Multivariate cox model Concordance (C-index): 0.683. Likelihood ratio test: χ² = 38.99, p < 0.001 Modified Poisson regression models with robust standard errors were used to evaluate associations between baseline clinical variables and regression outcomes. In multivariable analysis, low-grade cytology (ASCUS/LSIL) and non-16/18 HPV genotypes were associated with an increased likelihood of regression, with risk ratios (RR) of 1.30 (95% CI 1.06–1.59; p = 0.011) and 1.28 (95% CI 1.04–1.57; p = 0.019), respectively. Lesion extent did not show a significant association with regression outcomes. (Table 4) Table 4Modified Poisson regression model with robust standard errors for regression outcomesVariableRisk Ratio (RR)95% CIp-valueCytology (ASCUS/LSIL vs. ASC-H/HSIL)1.301.06–1.590.011HPV genotype (non-16/18 vs. HPV16/18)1.281.04–1.570.019 Modified Poisson regression model with robust standard errors for regression outcomes Analyses focused on progression outcomes confirmed that high-grade cytology (ASC-H/HSIL) and HPV16/18 positivity were associated with an increased likelihood of progression, whereas lesion extent was not independently associated with progression in multivariable models. (Table 5) Table 5Modified Poisson regression model with robust standard errors for progression outcomesVariableRR95% CIp-valueCytology (ASC-H/HSIL vs. ASCUS/LSIL)2.171.19–4.000.012HPV (16/18 vs. non-16/18)1.961.09–3.560.025Lesion extent (3–4 vs. 1–2 quadrants)1.310.72–2.410.37 Modified Poisson regression model with robust standard errors for progression outcomes Overall, these findings suggest that the clinical evolution of CIN2 during active surveillance is more strongly associated with virological and cytological factors than with the initial colposcopic extent of the lesion. Discussion This study provides a detailed analysis of the clinical outcomes of CIN2 lesions in a cohort of HPV-positive women undergoing active surveillance, with a follow-up of up to 48 months. Using survival analyses (Kaplan-Meier), modified Poisson regression models with robust standard errors, and Cox models, we were able to accurately assess the cumulative probability of regression and progression, as well as the impact of key clinical and virologic covariates. The results offer important insights for optimizing the conservative management of CIN2 and are part of the broader international debate on the safety of active surveillance. The overall regression rate observed (61%) is consistent with that reported in the meta-analysis by Tainio et al. [2], which documents spontaneous regression in 50–60% of cases within 24 months, especially in young women. Regression was defined by two consecutive negative co-tests without systematic histological confirmation in all cases. This approach reflects routine clinical surveillance practice and is supported by the high negative predictive value of combined cytology and HPV testing reported in the literature. However, the temporal analysis of the Kaplan–Meier curves in our study provides additional relevant information: regression is mainly concentrated in the first 24–36 months, while beyond this interval the cumulative probability of further regression tends to stabilize. This trend suggests that CIN2 persisting beyond three years may represent a lesion biologically less prone to spontaneous resolution and potentially more stable from a molecular point of view. The rate of persistence at 48 months (13.8%) is lower than other studies reporting persistence rates of 20–30% at 24 months [2–10], suggesting the effectiveness of active surveillance in the initial selection of patients. However, it should be considered that persistence beyond 36 months, as observed in 36 patients in our study, has been associated in the literature with a significantly increased risk of progression to CIN3 or microinvasion [2, 17]. Although not classifiable as progression, these patients remain exposed to increased risk of progression and raise questions about the optimal duration of active surveillance. While time-to-event analyses were performed for regression, progression to CIN3 was analyzed as a clinical endpoint, given the structure of the dataset. Histological progression to CIN3 occurred in 25.5% of patients during follow-up. This figure appears slightly higher than that reported in similar studies, which indicate progression rates between 10% and 20% [2, 10]. A possible explanation for this discrepancy lies in the composition of the cohort: in our sample, there was a higher prevalence of HPV 16/18 infections (40.2%) and initial ASC-H/HSIL cytology (38%), two factors known to have oncogenic potential. Furthermore, the higher progression to CIN3 + in our study was observed during the 48-month follow-up, a longer time window than that of many published cohorts, which are often limited to 24–36 months [2]. In addition, the relatively smaller sample size compared to large population-based studies may have contributed to variability in the estimated progression rates. This suggests that cumulative progression risk may be underestimated at shorter follow-ups, especially in the presence of unfavorable prognostic factors. Covariate analysis using modified Poisson regression models with robust standard errors and Cox models confirmed that high-grade cytology and the presence of HPV 16/18 were significantly associated with an increased risk of progression. This biological behavior confirms the greater oncogenicity of HPV 16 and 18, genotypes historically associated with longer viral persistence and a higher risk of progression to CIN3 or carcinoma in situ [13, 16]. These data are consistent with those reported by Schiffman et al. and Egemen et al., who identified these parameters as high-risk markers in the management models proposed by the American Society for Colposcopy and Cervical Pathology (ASCCP) [11, 12]. Baseline cytology also plays a more prominent role than histology in the risk stratification process. In some cases, cervical biopsy may underestimate the true severity of the lesion, especially when the dysplastic areas are multifocal or not fully accessible during colposcopic examination. In this context, the presence of HSIL cytology takes on a relevant diagnostic role, as it may reflect the presence of areas of severe dysplasia that were not adequately sampled in the initial biopsy. This phenomenon, also described in the ASCUS-LSIL Triage Study (ALTS study), highlights how a more severe cytology than histology should not be underestimated, but rather considered a sign of possible underestimation of the lesion and the presence of inadequately sampled high-grade disease [17]. Consequently, in patients with HSIL cytology, the conservative approach requires particular caution. The probability that a histological CIN2 represents a diagnostic underestimation of an underlying CIN3 is in fact significantly higher [18, 19]. These data are consistent with what is reported in the literature: classic studies and more recent guidelines have highlighted how high-grade cytological alterations reflect an increased risk of underlying histological lesion and a lower probability of spontaneous regression, especially in cases associated with high-risk HPV infections [20, 21]. Our study should be interpreted as providing confirmatory clinical evidence supporting risk stratification in CIN2 management, rather than as a mechanistic investigation. These findings reinforce the concept that histological diagnosis of CIN2 alone is not sufficient to determine eligibility for active surveillance. Instead, an integrated approach is needed that considers HPV genotype and cytology for a true individual risk stratification. This approach is consistent with the 2019 ASCCP guidelines, which propose management based on cumulative risk rather than individual histological categories [1]. Our findings suggest that HPV genotype and baseline cytology may contribute to risk assessment and support patient counseling during active surveillance of CIN2. Consistently, the 2023 recommendations from the European Society of Gynaecological Oncology (ESGO), the European Society for Radiotherapy and Oncology (ESTRO), and the European Society of Pathology (ESP) also emphasize HPV16/18 persistence as an indication for closer follow-up, regardless of histological stability [22, 23]. Our findings also fit into the broader debate on long-term cancer risk, while not allowing a direct estimation of such risk. Rather than directly estimating long-term cancer risk, our study identifies factors associated with progression during medium-term follow-up. While HPV genotype and cytology are well-established risk factors, our study adds clinically relevant evidence by quantifying their impact on medium-term outcomes in a strictly defined surveillance cohort. It is important to underline that several cohort studies on large European populations have highlighted a higher 20-year cumulative risk of invasive cancer in women managed with surveillance alone compared to those treated with LEEP [10, 24]. Similar data have been reported by Danish and British studies, which observed a 5–10-fold increased long-term risk in case of persistent or HPV16 + lesions, even after apparent cytological regression [25, 26]. It is essential to clarify that our study does not aim to directly estimate the incidence of cervical cancer, but rather to identify, within a medium-term follow-up, biological and clinical profiles that the literature recognizes as being associated with long-term cancer risk. In this context, our findings may help to explain which CIN2 subgroups could contribute to the increased risk observed in long-term population-based studies. However, progression to CIN3 + represents a clinically significant endpoint and is widely considered a precursor of invasive disease. It should be interpreted as a clinically relevant intermediate endpoint and a surrogate marker of increased future cancer risk, rather than as a direct measure of oncologic outcomes. The identification of clinical and virological factors associated with this progression therefore allows the early recognition of biological profiles potentially linked to an increased risk of progression over time. In our cohort, persistence was observed in 36 patients (13.8%), whereas progression to CIN3 occurred in 70 patients (25.5%) during follow-up. These percentages, if projected over longer follow-up, could translate into an increased risk of invasive cancer, as confirmed in the literature, where women with persistent HSIL have a cumulative risk of cancer up to 5–10 times higher than those treated with excision. Although excisional treatment reduces the immediate risk, a residual risk of distant cancer (0.3–0.8% at 20 years) has also been documented in women undergoing LEEP, suggesting that neither surveillance nor treatment completely eliminates long-term risk. However, in high-risk patients (HSIL, HPV16-positive), conservative management could delay the treatment of lesions that may progress, potentially increasing the likelihood of progression over time. These data support the need for a personalized decision-making model, in which active surveillance is reserved for patients with favorable profiles (ASC-US/LSIL cytology, HPV no. 16/18 positive), while in patients with unfavorable prognostic factors, early excision may be the safest choice, reducing the cumulative risk of invasive cancer. These findings do not introduce novel risk factors, but rather provide confirmatory clinical evidence supporting the prognostic relevance of established parameters—such as HPV genotype and cytology—in the context of active surveillance of CIN2. In recent years, major international scientific societies have progressively defined more rigorous criteria for active observation of CIN2, recommending careful patient selection, strict adherence to follow-up protocols, clear treatment thresholds, and surveillance exit criteria [1, 22, 23]. The reported C-index should be interpreted cautiously, as the model was not developed or validated as a formal predictive tool. Therefore, the multivariable analyses presented in this study should be interpreted as exploratory association models requiring confirmation in independent prospective cohorts. The limitations of this study deserve consideration. The retrospective design, the selective inclusion criteria, and the high level of adherence to follow-up may limit the generalizability of these findings, particularly to real-world screening populations where adherence and diagnostic conditions are more variable. The sample size does not allow for high statistical power for rare events such as invasive carcinoma. The relatively short follow-up does not allow the assessment of outcomes beyond 5–10 years. Variables such as lesion depth, cone volume (if treated), or additional biomarkers (HPV E6–E7 mRNA, Ki67, HPV methylation) were not integrated. Post-excisional histological outcomes were not analyzed, as they were beyond the scope of our study [27–30]. The cohort is selective (p16 + diagnosis, high adherence to follow-up), which limits generalizability to populations with lower compliance. In addition, histological confirmation was not systematically performed in all cases of apparent regression; therefore, a potential overestimation of regression due to sampling limitations cannot be entirely excluded. Regression was defined by two consecutive negative co-tests without systematic histological confirmation in all cases. This approach reflects routine clinical surveillance practice and is supported by the high negative predictive value of combined cytology and HPV testing reported in the literature [11–13]. The use of additional biopsies in cases of cytological or colposcopic worsening may have introduced detection bias and contributed to an overestimation of progression rates in higher-risk subgroups. Patients with high-risk features (HSIL cytology or HPV16/18 positivity) were more likely to undergo additional biopsies in response to worsening findings during follow-up. This may have introduced a verification bias, potentially increasing the probability of detecting progression in higher-risk subgroups. ASC-H and HSIL were grouped as high-grade cytological abnormalities; however, these categories have different prognostic values for CIN3+, with ASC-H generally showing a lower positive prognostic value, potentially introducing some degree of heterogeneity within this group. Information on HPV vaccination status was not systematically available and was therefore not included in the analysis; thus, its potential impact on lesion regression or progression could not be evaluated. Because lesion status was assessed at scheduled follow-up visits rather than continuously, event times should be considered interval-censored. Although Cox proportional hazards models are commonly used in this setting, discrete-time survival approaches may represent an alternative analytical strategy in future studies. The regression analyses were performed using modified Poisson regression models with robust standard errors to directly estimate risk ratios. However, these multivariable models should be interpreted as exploratory association analyses and require confirmation in larger independent cohorts. Strengths of this study include diagnostic homogeneity (all p16+), the integration of cytological, virological, and colposcopic parameters, and a well-defined surveillance protocol. This work contributes to the literature by providing real-world clinical data that can support personalized decision-making. Our study was not designed to investigate the biological mechanisms underlying CIN2 progression or regression; rather, it provides confirmatory clinical evidence supporting risk stratification based on virological and cytological factors. Further studies integrating molecular biomarkers—such as HPV E6/E7 mRNA expression, host–virus interaction markers, and epigenetic profiling—are needed to better elucidate the biological pathways driving lesion evolution. The regression models were not internally validated using calibration metrics or bootstrapping; therefore, their prognostic performance should be considered exploratory and requires confirmation in larger, independent cohorts. Conclusions Our study contributes to the ongoing debate on the conservative management of CIN2, providing real-world evidence that supports the risk-based approach proposed by the 2019 ASCCP guidelines (cytology, HPV status, persistence). Active surveillance of CIN2 may not be equally appropriate across all patient subgroups. Although effective within a follow-up period of up to 48 months in a substantial proportion of patients, the presence of biological and virological risk factors identifies subgroups in which conservative management could, over time, be associated with an increased risk of progression. Acknowledgements Not applicable. Abbreviations Cervical Intraepithelial Neoplasia grade 2 Cervical Intraepithelial Neoplasia grade 3 Human Papillomavirus High-grade Squamous Intraepithelial Lesion Low-grade Squamous Intraepithelial Lesion Authors' contributions MTB, AP, GV.: conceptualized the data. TA, EP, AGC, MMP: curated the data. Formal analysis by MT. MTB wrote the original draft and edited the draft. All authors read and approved the final manuscript. Funding This research received no external funding. Data availability All data supporting the findings of this study are available within the paper. Declarations Ethics approval and consent to participate This study conforms to the principles of the Declaration of Helsinki, as revised in 2013. Based on Italian law, the University Hospital’s ethics committee (Catania 2), Azienda Ospedaliero-Universitaria Policlinico Rodolico–San Marco, Catania, Italy, waived the requirement for ethical approval and informed consent because the study used previously archived data, according to current legislation (20 March 2008) (AIFA): https://www.gazzettaufficiale.it/eli/id/2008/03/31/08A02109 (accessed on 12 December 2025). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data supporting the findings of this study are available within the paper.