In our current cohort of 334 women positive for HPV 16/18, we assessed the clinical performance of PAX1/JAM3 methylation across different grades of cervical lesions, and its effectiveness in triaging diagnosis. The methylation detection exhibited a good balance of sensitivity and specificity in CIN2 and CIN3 lesions, achieving AUC values of 0.921 and 0.905, respectively, outperforming cytology alone or cytology combined with methylation tests. Importantly, triaging using PAX1/JAM3 methylation in HPV16/18-positive women can markedly reduce excessive colposcopy referrals without increasing the risk of missing high-grade lesions.

Given the high risk of HPV16/18 infections progressing to precancerous lesions, as evidenced by a 10-year follow-up study which found a progression rate of 20.7% for women infected with HPV16 and 17.7% for those infected with HPV18 [25], it is recommended that individuals who test positive for HPV 16 or 18 should be directly referred for colposcopy [7,8,9,10]. However, approximately 70% of HPV16/18-positive women in this study showed no lesions or only CIN1, consistent with findings from some larger cohort studies [26,27,28], suggesting that many women may not require immediate colposcopy referral. Using methylation tests to triage these populations can effectively reduce colposcopy referral rates. This approach is supported by previous reports on other gene methylation tests in hrHPV-positive women, such as WID-qCIN [29], FAM19A4/miR124-2 [30], ZNF671/ASTN1/ITGA4/RXFP3/SOX17/DLX1 [31], and PAX1/SOX1 [32]. Particularly notable is the performance of PAX1/JAM3 methylation in our study, demonstrating 89.0% sensitivity and 95.3% specificity in detecting CIN2+ lesions.

The clinical significance of the methylation of PAX1 and JAM3 in cervical cancer screening has been previously reported [19, 33,34,35]. PAX1 gene expression can epigenetically activate a series of phosphatases, which subsequently suppress signaling pathways such as EGF/MAPK, thereby inhibiting malignant phenotypes. HPV infection can lead to high methylation of the PAX1 gene, resulting in downregulation or loss of its expression [36, 37]. JAM3 is involved in the formation of tight junctions between cells, facilitating the regulation of vascular permeability and the migration of leukocytes across endothelial surfaces [38]. Combining PAX1 and JAM3 into a methylation panel has demonstrated promising efficacy in studies involving patients with persistent HPV infection and self-sampling methods [21, 23, 39]. In this cohort, we highlight its triage value in HPV 16/18-positive women. Assuming that a positive result would trigger a colposcopy referral, an average of only 1.12 referrals would be needed to detect one CIN2+ case. The findings of the study suggest that HPV16/18-positive women who test negative for PAX1m/JAM3m may not immediately require referral to colposcopy, especially young women.

As women age, there is a decline in hormone levels which leads to gradual changes in the cervix. This leads to the inward movement of the transformation zone into the cervical canal and thinning of the cell layer, which poses challenges for the accuracy of cytology and colposcopy in detecting CIN2+ lesions. Perimenopausal or postmenopausal women experience reduced natural clearance of HPV, making detection of HPV 16/18 infection in this age group of heightened clinical concern [40, 41]. Colposcopy poses challenges due to atrophy, retraction, and limited visualization of the transformation zone, potentially leading to misdiagnosis in cytology and biopsy [42]. Guidelines in some countries suggest considering diagnostic endocervical curettage (ECC) or large loop excision of the transformation zone (LLETZ) in women with abnormal screening results when visualization of the transformation zone is incomplete [43, 44]. LLETZ and excessive ECC not only increase the financial burden on patients, but also pose a higher risk of post-operative complications such as adhesions, pain, cervical stenosis, and others. Additionally, they decrease the likelihood of patients attending follow-up appointments. The CISCER triage strategy has demonstrated the ability to effectively identify high-risk patients without over-treating and to reduce missed diagnoses, particularly in women aged 45 and older. The promising efficiency of PAX1/JAM3 methylation detection in our study suggests it could serve as a valuable risk marker to guide clinical management of older women, independent of cytological results.

Following CISCER triage, the colposcopy referral rate was lowest in the youngest age group, and the incidence of false-positive PAX1/JAM3 methylation was lowest among these women. Regression rates for CIN2 are notably high in young women (< 30 years), reaching estimates of up to 70% at 3 years [45, 46]. Hence, women at this age are at higher risk of over-referral, which can be effectively minimized by methylation triage. A recent study demonstrated a marked decline in methylation positivity rates with decreasing age [47]. In our research, we observed lower PAX1/JAM3 methylation levels in younger women. The difference in methylation levels by age may be correlate with spontaneous CIN2 regression rate and the duration of HPV infection [23]. Women with lower methylation levels may be less prone to progressing to higher-grade lesions [48]. For young women with higher HPV infection rates, CISCER triage indeed can alleviate potential anxiety and unnecessary surgical treatments. In China, where fertility rates have declined in recent years, more clinical data could be gathered on methylation testing as a triage method for HPV 16/18-positive young women with CIN2 lesions in the future.

All of the cervical cancer patients were identified through our methylation tests. The methylation of the PAX1 and JAM3 genes showed high consistency in risk stratification of women infected with HPV16/18. However, we observed that PAX1m was negative in one adenocarcinoma patient, while the patient was not misdiagnosed due to positive result of JAM3m. The incidence of cervical adenocarcinoma is lower than that of cervical squamous cell carcinoma, but its occurrence has been gradually rising in recent years. It usually originates within the cervical canal, characterized by multifocal or skip lesions. Conventional clinical examination methods have limited sensitivity, thereby increasing the risk of missed diagnoses and misdiagnosis [11, 49,50,51]. With advancements in molecular technology, an increasing number of markers and commercial tests are available for cancer detection. Nonetheless, it remains elusive whether certain markers are more closely associated with specific types of cancer. In future studies, we will expand our sample size of adenocarcinoma patients to more comprehensively evaluate the relationship between JAM3 methylation and the development of adenocarcinoma.

In this study, PAX1m/JAM3m showed superior performance compared to cytology, and the combined effect of both was not significantly better than that of PAX1m/JAM3m alone. Higher levels of methylation directly correspond to more advanced cervical disease [48], while the cytology results rely on pathologists’ interpretation [52]. The 8th National Congress of Colposcopy and Cervical Pathology (CSCCP) indicated that ASC-US/LSIL accounts for about 90% of cytological abnormalities in China. However, the incidence of high-grade lesions among these cases is generally below 30% [53,54,55,56], as observed in this study. Due to the low specificity of Cytology ≥ ASC-US/LSIL and the low sensitivity of Cytology ≥ HSIL, combining cytology with CISCER does not significantly enhance the triage performance.

With the WHO's strategy to eliminate cervical cancer introduced [57], prophylactic HPV vaccines have been recommended as a pivotal preventive measure. Their adoption has proven effective in reducing disease rates and the prevalence of HPV 16/18 [58, 59]. Despite this, HPV vaccine coverage remains notably low in China. The cumulative vaccination rate among women aged 9–45 was estimated to be less than 3% between 2018 and 2020 [60]. Therefore, cervical screening is still necessary. Early diagnosis and management strategies for women infected with hrHPV, especially HPV16/18, are currently focal points deserving significant attention. We propose that PAX1/JAM3 methylation triage strategies can be instrumental in addressing these challenges.

There are also some limitations in this study. Histopathological results from biopsies, which may not always be conducted at the punctum maximum rather than through surgical procedures, could potentially introduce some bias into the outcomes of this study. Additionally, as a real-world study, not all referred colposcopy patients underwent biopsies. This may cause a clinical selection/bias that could affect the performance of the methylation test and the accuracy of immediate CIN2/CIN3+ risk assessment. Future directions will involve expanding the sample size of patients with AD and conducting more rigorous clinical trials to evaluate the clinical value of methylation testing.