ABSTRACT
Objective
To establish a better triage strategy using SOX1/PAX1 methylation detection for high‐risk HPV (hrHPV)‐positive women than cytology.
Design
A cohort study.
Setting
Population‐based cervical cancer (CC) screening cohort.
Population
A total of 5684 women were enrolled.
Methods
SOX1/PAX1 methylation was detected by quantitative methylation‐specific PCR using cytologic residue from hrHPV‐positive women at baseline in a 3‐year CC screening cohort. Risk stratification ability was evaluated by the immediate and cumulative cervical intraepithelial neoplasia (CIN) grade 2/3 or worse (CIN2+/3+) risks.
Main Outcome Measures
CIN3+ and CIN2+.
Results
At baseline, 682 hrHPV‐positive women were included with 63 CIN2+ and 39 CIN3+. Over 3 years, 109 CIN2+ and 62 CIN3+ were detected. Methylation demonstrated better risk stratification than cytology among hrHPV‐positive women. When compared with current practice triage strategy (Strategy A), post hoc re‐triage analysis showed that methylation triage for all hrHPV‐positive women (Strategy C) significantly increased sensitivity (97.44%/83.87% vs. 84.62%/64.52%, p = 0.0476/0.014), specificity (83.36%/85.00% vs. 73.41%/73.55%, p = < 0.001/< 0.001), positive predictive value (26.21%/35.86% vs. 16.18%/19.61%, p = 0.022/0.001), and negative predictive value (99.81%/98.14% vs. 98.74%/95.40%, p = 0.040/0.012) in detecting immediate and 3‐year cumulative CIN3+. Similar improvements were observed for methylation triaged for HR12‐positive (Strategy B) and for combined HPV and cytology primary screening (Strategy D). More importantly, all methylation‐based triage strategies reduced colposcopies and postponed follow‐up intervals compared to Strategy A over the 3‐year period of CC screening.
Conclusion
SOX1/PAX1 methylation showed better risk stratification compared with cytology and enabled more efficient management of HPV‐positive women, though external validation and head‐to‐head comparisons with other triage methods are needed.
Keywords: cervical cancer, risk stratification, SOX1/PAX1 methylation, triage efficacy
1. Introduction
Cervical cancer (CC) is currently the fourth most common cancer in women worldwide [1]. CC screening could reduce CC risk by detecting and treating precancerous lesions. A number of international guidelines recommended primary human papillomavirus (HPV) screening as one of the main CC screening strategies, due to its high sensitivity and low long‐term risk of CC provided by negative high‐risk HPV (hrHPV) [2, 3, 4]. However, it would identify more than 10% of HPV‐positive women in the screening population, and less than 1% would lead to CC [5]. Therefore, triage tests for women with hrHPV‐positive results are needed to further distinguish the transient infection from cervical lesions. Traditional cytology triage could not avoid its subjectivity and low sensitivity for detecting dysplastic lesions, which would lead to under‐diagnosis [6, 7, 8]. Moreover, once positive HPV is detected, it inevitably causes anxiety and even over‐treatment [9]. Thus, it is necessary to explore an additional triage approach, which could identify as many precancerous cervical lesions as possible and avoid unnecessary referrals for colposcopy.
DNA promoter hypermethylation of various tumour suppressor genes has been recognised as an important epigenetic regulation factor involved in the tumorgenesis and progression for CC [10, 11, 12]. Several studies reported that methylation of CpG islands in the PAX1 and SOX1 promoter region could regulate the occurrence of cervical neoplasia [11, 13, 14]. Zhang et al. found PAX1 overexpression restrained proliferation, migration, and improved cisplatin sensitivity by interfering with the WNT/TIMELESS axis in CC cells [15]. Lin et al. also found that SOX1 inhibited CC cells' invasion through a β‐catenin‐independent pathway by upregulating CDH1 and downregulating SLUG [16]. These findings indicated that PAX1 and SOX1 methylation might play as potential biomarkers for the identification of persistent hrHPV infection, cervical intraepithelial lesions and cancer [17, 18, 19]. PAX1 DNA methylation has been reported to demonstrate good performance in detecting cervical intraepithelial neoplasia (CIN) grade 3 or worse (CIN3+), achieving sensitivities ranging from 64%–74%, alongside specificities ranging from 66.86%–91.60% [20, 21]. SOX1 also showed a promising predictive performance for CIN3+ with sensitivity of 71% and specificity of 77% in a hospital‐based case control study [21]. However, data from population‐based cohort studies evaluating the combined SOX1/PAX1 methylation panel as predictors of cervical precancer and cancer remain limited. We aimed to evaluate the risk stratification performance of SOX1/PAX1 methylation as a triage test for hrHPV‐positive women and to compare several methylation‐based triage strategies with current cytology‐based approaches over a 3‐year follow‐up.
2. Methods
2.1. Study Design and Population
The study was nested within our previously published population‐based 3‐year prospective multicenter CC screening cohort [22]. Which comprised 8370 eligible women conducted in Jinyun, Jingning, and Fuyang cities (Zhejiang Province) and Xinmi city (Henan Province). In the present study, participants from Xinmi were excluded due to unavailability of cytology residue, resulting in an analytic multicenter cohort of 5684 women. In brief, women received HPV and cervical cytology detection at baseline (V0), and any of the following situations would be referred for colposcopy: (1) positive HPV16/18, (2) atypical squamous cells of undetermined significance or worse (≥ ASC‐US) cytology. Women with any positive HPV or ≥ ASC‐US cytology at V0 were recalled for cytology detection at the first and second year visit (V1–V2), and those with ≥ ASC‐US cytology during follow‐up would be referred for colposcopy. At the third year visit (V3), all enrolled women were recalled for HPV and cytology co‐testing, and women with any positive HPV or ≥ ASC‐US cytology results were referred to colposcopy. Cytology samples collected from hrHPV‐positive women at V0 that contained enough residue were further tested for methylation. The primary and second endpoints were CIN3+ and CIN grade 2 or worse (CIN2+), respectively. Present study was approved by the Ethics Committee of Women's Hospital, School of Medicine, Zhejiang University (approval number: IRB‐20240112‐R).
2.2. HPV and Cervical Cytology Detection
All subjects were detected by ThinPrep cytologic testing (Hologic, USA) and HPV genotyping kit (Jiangsu Bioperfectus Technologies Ltd.), including 14 hrHPV (HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and 68), 4 intermediate‐risk (HPV 26, 53, 73, and 82), and 3 low‐risk (HPV 6, 11, and 81) HPV types. Women with 14 hrHPV infection were included in the current study. Cytology and HPV were independently detected and blinded to each other. Histologic evaluations were further blinded to both cytologic and HPV results. Histopathologists were unaware of methylation status because methylation testing was performed post hoc. P16 immunohistochemistry was only performed selectively for clinical management purposes in cases diagnosed as high‐grade squamous intraepithelial lesion (HSIL)/CIN2 to aid treatment decisions. Cytologic results were classified by the Bethesda System [23], including NILM, ASC‐US, low‐grade squamous intraepithelial lesion (LSIL), and ≥ HSIL (included ASC‐H, HSIL, AGC, and cancer). The histologic classification included normal epithelium, LSIL/CIN1, HSIL/CIN2, HSIL/CIN3, adenocarcinoma in situ, and cancer.
2.3. Methylation Testing
Genomic DNA was extracted from cytologic residue using the QIAamp DNA Mini Kit (Qiagen GmbH, Hilden, Germany) following the manufacturer's recommendations. The genomic DNA underwent bisulfite conversion using the EZ DNA Methylation‐GoldM Kit (Zymo Research, CA, USA) as instructed by the manufacturer's manual. Quantitative methylation‐specific PCR was performed on the Applied Biosystems 7500 instrument (Applied Biosystems, CA, USA) with the SOX1 and PAX1 gene methylation detection kit (Hybribro Pharmaceutical Technology Co. Ltd., Guangzhou, China), following the manufacturer's instructions. The SOX1/PAX1 methylation assays were blinded to HPV, cytologic and histologic results. Ct values for SOX1, PAX1, and β‐actin (as internal reference) were obtained for each sample; Ct value of β‐actin between 15 and 35 would be considered as qualified. Ct values below the cut‐off of either SOX1 (38.6) or PAX1 (38) were considered positive methylation status.
2.4. Statistical Analysis
Description statistics were calculated to assess the distribution of the 14 hrHPV genotypes and their CIN2+/3+ risks. The 95% confidence intervals (95% CI) for the CIN2+/3+ risks were calculated using the Clopper‐Pearson exact method. Four screening strategies were compared, including traditional HPV primary screening and cytology triage for the other 12 hrHPV (HR12) (Strategy A), HPV primary screening and methylation triage for HR12 (Strategy B), methylation triage for all hrHPV‐positive women (Strategy C), traditional HPV/cytology combined primary screening and further risk‐based triage using methylation (Strategy D). Strategies A‐D were all modelled using data from the same screening cohort, and strategies B‐D were post hoc re‐triage studies. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) and their 95% CI were utilised to estimate the clinical performance of different triage strategies. Colposcopy/CIN was utilised to describe the number of colposcopies per detected CIN2+/3+ case, while Visit/CIN was utilised to describe the number of visits per detected CIN2+/3+ case over 3 years. Sensitivity and specificity were compared using McNemar's test for paired proportions, whereas comparisons of PPV, NPV, Colposcopy/CIN, and Visit/CIN were performed with Chi‐square tests. p < 0.05 was recognised as statistically significant. All statistical analyses were conducted using SAS 9.4.
3. Results
3.1. Better Risk Stratification Ability of Methylation in hrHPV‐Positive Women
Among 5684 women enrolled in the cohort, 734 (12.91%) women tested hrHPV‐positive at V0. Finally, 682 women were included for analysis after excluding those without enough cytologic residue or valid DNA methylation results (Figure 1). The median age of hrHPV‐positive women was 49 years (range, 23–65 years). At V0, 95 (13.93%) cases were HPV 16/18‐positive, 134 (19.65%) were ≥ ASC‐US, and 145 (21.26%) were positive for methylation (Table S1). At baseline, 63 women had CIN2+, including 39 with CIN3+. Over 3 years, 109 women had CIN2+, including 62 with CIN3+ (including 1 cervical squamous cell carcinoma).
FIGURE 1.
The flowchart of current study.
As shown in Table S1, SOX1/PAX1 methylation positive rate increased with increasing grades of cytology and CIN. And cytology results could effectively stratify HPV‐positive women into high‐risk (≥ ASC‐US cytology) and low‐risk (with no intraepithelial lesion or malignancy (NILM) cytology) groups (Figure 2 and Table S2). However, negative methylation identified significantly lower immediate CIN3+ and 3‐year cumulative CIN2+/3+ risks compared to NILM cytology (immediate CIN3+ risk: 0.19% vs. 1.82%; 3‐year cumulative CIN2+/3+ risks: 6.52%/1.86% vs. 10.22%/5.29%), with corresponding relative risks of 0.10 (95% CI: 0.01–0.74) for immediate CIN3+ detection and 0.64 (95% CI: 0.44–0.93)/0.35 (95% CI: 0.18–0.69) for 3‐year cumulative CIN2+/3+ risks. While positive methylation could identify higher risks (34.48%/26.21% vs. 34.33%/21.64% and 51.03%/35.86% vs. 39.55%/24.63%) compared with ≥ ASC‐US cytology, although without statistical significance (Figure 2 and Table S2). Similar results were also observed in both HPV 16/18‐positive and any HR12‐positive subgroups (Table S2). Thus, SOX1/PAX1 methylation exhibited better risk stratification than cytology among HPV‐positive women.
FIGURE 2.
Cumulative risks of CIN2+ and CIN3+ by methylation and cytology results. (A, B) the cumulative risk of CIN2+ (A) and CIN3+ (B) in HPV‐positive women; (C, D) the cumulative risk of CIN2+ (C) and CIN3+ (D) in HPV 16/18‐positive women; (E, F) the cumulative risk of CIN2+ (E) and CIN3+ (F) in other 12 hrHPV‐positive women.
3.2. The Efficacy of Methylation Triage Strategy for HR12‐Positive Women
When compared with cytologic triage for HR12‐positive women (strategy A), methylation triage (strategy B) showed significantly higher sensitivity in detecting immediate CIN3+ (97.44% vs. 84.62%), higher sensitivity and NPV in detecting 3‐year cumulative CIN2+/3+ (79.82%/85.48% vs. 64.22%/64.52% and 95.34%/98.09% vs. 91.84%/95.40%). While the specificity and PPV between strategy A and B were comparable (Table 1). In addition, strategy B could significantly decrease the number of colposcopy/3‐years CIN3+ when compared with strategy A (6.09 vs. 8.18, p = 0.024) (Table S3). These results suggested that SOX1/PAX1 methylation might be a potentially alternative triage method for HR12 positive women.
TABLE 1.
The screening efficacy of different triaged strategies by cytology and methylation for identifying CIN2+/3+ in hrHPV positive women.
| Women, No./Total No. (%) [95% CI] | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Strategy A | Strategy B | Strategy C | Strategy D | P (A vs. B) | P (A vs. C) | P (B vs. C) | P (A vs. D) | P (C vs. D) | |
| Baseline | |||||||||
| CIN2+ | |||||||||
| Sensitivity | 53/63, 84.13% (72.28–91.72) | 56/63, 88.89% (77.84–95.04) | 50/63, 79.37% (66.96–88.14) | 54/63, 85.71% (74.10–92.86) | 0.434 | 0.489 | 0.144 | 0.803 | 0.349 |
| Specificity | 468/619, 75.61% (71.99–78.90) | 465/619, 75.12% (71.48–78.44) | 524/619, 84.65% (81.51–87.35) | 514/619, 83.04% (79.79–85.86) | 0.843 | < 0.001 | < 0.001 | 0.001 | 0.442 |
| PPV | 53/204, 25.98% (20.22–32.66) | 56/210, 26.67% (20.93–33.28) | 50/145, 34.48% (26.92–42.88) | 54/159, 33.96% (26.77–41.95) | 0.874 | 0.086 | 0.114 | 0.098 | 0.924 |
| NPV | 468/478, 97.91% (96.06–98.93) | 465/472, 98.52% (96.83–99.35) | 524/537, 97.58% (95.79–98.65) | 514/523, 98.28% (96.64–99.16) | 0.479 | 0.725 | 0.286 | 0.667 | 0.424 |
| CIN3+ | |||||||||
| Sensitivity | 33/39, 84.62% (68.79–93.59) | 38/39, 97.44% (84.92–99.87) | 38/39, 97.44% (84.92–99.87) | 38/39, 97.44% (84.92–99.87) | 0.048 | 0.048 | 1.000 | 0.048 | 1.000 |
| Specificity | 472/643, 73.41% (69.78–76.75) | 471/643, 73.25% (69.62–76.60) | 536/643, 83.36% (80.20–86.11) | 522/643, 81.18% (77.90–84.09) | 0.950 | < 0.001 | < 0.001 | 0.001 | 0.307 |
| PPV | 33/204, 16.18% (11.55–22.12) | 38/210, 18.10% (13.27–24.12) | 38/145, 26.21% (19.42–34.28) | 38/159, 23.90% (17.66–31.43) | 0.605 | 0.022 | 0.067 | 0.066 | 0.643 |
| NPV | 472/478, 98.74% (97.15–99.49) | 471/472, 99.79% (98.64–99.99) | 536/537, 99.81% (98.80–99.99) | 522/523, 99.81% (98.77–99.99) | 0.060 | 0.040 | 0.927 | 0.044 | 0.985 |
| 3‐year cumulative | |||||||||
| CIN2+ | |||||||||
| Sensitivity | 70/109, 64.22% (54.41–73.01) | 87/109, 79.82% (70.82–86.66) | 74/109, 67.89% (58.17–76.33) | 78/109, 71.56% (61.99–79.59) | 0.010 | 0.567 | 0.045 | 0.246 | 0.555 |
| Specificity | 439/573, 76.61% (72.89–79.98) | 450/573, 78.53% (74.90–81.78) | 502/573, 87.61% (84.56–90.14) | 492/573, 85.86% (82.68–88.56) | 0.436 | < 0.001 | < 0.001 | < 0.001 | 0.384 |
| PPV | 70/204, 34.31% (27.91–41.32) | 87/210, 41.43% (34.75–48.43) | 74/145, 51.03% (42.64–59.37) | 78/159, 49.06% (41.09–57.06) | 0.136 | 0.002 | 0.074 | 0.005 | 0.730 |
| NPV | 439/478, 91.84% (88.92–94.06) | 450/472, 95.34% (92.92–96.99) | 502/537, 93.48% (90.96–95.36) | 492/523, 94.07% (91.60–95.87) | 0.028 | 0.315 | 0.202 | 0.167 | 0.691 |
| CIN3+ | |||||||||
| Sensitivity | 40/62, 64.52% (51.27–75.96) | 53/62, 85.48% (73.72–92.75) | 52/62, 83.87% (71.87–91.59) | 52/62, 83.87% (71.87–91.59) | 0.007 | 0.014 | 0.803 | 0.014 | 1.000 |
| Specificity | 456/620, 73.55% (69.86–76.94) | 463/620, 74.68% (71.03–78.02) | 527/620, 85.00% (81.89–87.67) | 513/620, 82.74% (79.48–85.59) | 0.6502 | < 0.001 | < 0.001 | < 0.001 | 0.280 |
| PPV | 40/204, 19.61% (14.53–25.86) | 53/210, 25.24% (19.63–31.77) | 52/145, 35.86% (28.19–44.29) | 52/159, 32.70% (25.61–40.65) | 0.1700 | 0.001 | 0.031 | 0.004 | 0.562 |
| NPV | 456/478, 95.40% (93.01–97.03) | 463/472, 98.09% (96.28–99.07) | 527/537, 98.14% (96.49–99.05) | 513/523, 98.09% (96.39–99.03) | 0.019 | 0.013 | 0.959 | 0.016 | 0.952 |
Note: Strategy A: primary HPV screening triaged by HPV 16/18 and cytology for other 12 hrHPV (HR12); Strategy B: primary HPV screening triaged by HPV 16/18 and methylation for HR12; Strategy C: primary HPV screening triage by methylation for all HPV‐positive women; Strategy D: HPV and cytology co‐testing screening triaged for all HPV‐positive women by methylation integrating with CIN2+/3+ risk thresholds.
3.3. The Efficacy of Optimised Methylation Triage Strategy for hrHPV‐Positive Women
Given the good performance of risk stratification by SOX1/PAX1 methylation in both HPV 16/18‐positive and HR12‐positive women, we further tried to use methylation as the triage tool for all HPV‐positive women (strategy C). Compared with strategy A, strategy C could significantly improve the sensitivity (97.44%/83.87% vs. 84.62%/64.52%, p = 0.048/0.014), specificity (83.36%/85.00% vs. 73.41%/73.55%, p < 0.001/< 0.001), PPV (26.21%/35.86% vs. 16.18%/19.61%, p = 0.022/0.001), and NPV (99.81%/98.14% vs. 98.74%/95.40%, p = 0.040/0.013) for immediate/3‐year cumulative CIN3+ (Table 1), while substantially reducing the number of colposcopy/immediate CIN3+ (3.82 vs. 6.18, p = 0.022) and colposcopy/3‐years CIN2+/3+ (3.17/4.70 vs. 4.85/8.18, p < 0.001) (Table S3). In addition, strategy C also demonstrated markedly increased specificity compared to strategy B in the detection of CIN2+/3+ (Table 1). These results suggested that SOX1/PAX1 methylation might be promising triage methods for all HPV infections including HPV16/18.
3.4. The Efficacy of Optimised Methylation Triage Strategy in hrHPV and Cytology Combined CC Screening
Considering HPV and cytology combined screening strategy is commonly used in real world scenario, we further assessed risk stratification ability of SOX1/PAX1 methylation for HPV‐positive women with different cytology results. As shown in Table S4, positive methylation in women with HR12‐positive and NILM cytology carried higher immediate and 3‐year cumulative CIN3+ risks (6.85% and 17.81%) compared with those who had negative methylation results (0.25% and 2.22%). So we further established a new strategy (strategy D) through integrating methylation status into combined CC screening, and different management strategies were suggested according to CIN2+/3+ risk thresholds. Although the ASCCP guideline has provided management risk thresholds of CIN3+, the risk threshold for CIN2+ was lack [3, 24], so we employed the 1 year follow‐up risk threshold of CIN2+ (> 6.2%) recommended by previous study [25].
Accordingly, subgroup with the immediate and cumulative risk during follow‐up of CIN3+ ≥ 4% or CIN2+ > 6.2% would be referred for colposcopy for strategy D (Table S4). In detail, the cumulative risks of CIN2+/3+ among HR12‐positive with NILM/ASC‐US cytology and negative methylation were all < 6.2%/4% during 3‐year follow‐up, indicating that a 3‐year return for routine retesting is acceptable. HR12‐positive with LSIL cytology and negative methylation is recommended to return in 2 years, due to the cumulative CIN2+ risk of 7.41% at V2. HPV16/18 positive with NILM cytology and negative methylation is recommended to follow‐up in 1 year due to the cumulative CIN2+ risk of 8.93% at V1. HPV16/18 positive with LSIL cytology and negative methylation was at very low risk, but a 1‐year return is still recommended due to the small sample size and high cancer risk associated with HPV16/18.
Compared with strategy A, strategy D could significantly improve the specificities of immediate and 3‐year cumulative CIN2+/3+ (83.04%/81.18% vs. 75.61%/73.41% and 85.86%/82.74% vs. 76.61%/73.55%, p = 0.001/0.001 and p < 0.001/< 0.001), and increased the sensitivities and NPVs of immediate and 3‐year cumulative CIN3+ (Table 1), which indicated SOX1/PAX1 methylation yielded better risk stratification for HPV‐positive women with normal or mild abnormal cytology results. In addition, strategies C and D demonstrate similar screening performance, but can be applied in different clinical scenarios for HPV primary screening and combined screening approaches, respectively. Further systematic estimates of the Colposcopy/CIN and Visit/CIN found strategy B, C, and D all significantly reduced the Colposcopy/CIN and Visit/CIN compared with strategy A (Table S3). These results showed that methylation‐based triage strategies could improve screening efficiency than cytology‐based triage methods.
4. Discussion
4.1. Main Findings
Cytologic testing has inherent shortcomings of low sensitivity and subjectivity in CC screening, and a variety of DNA methylation markers were considered to have the potential for CC screening. Thus, the present study evaluated the long‐term CIN2+/3+ predictive performance of SOX1/PAX1 methylation detection for the first time in HPV‐positive women through a 3‐year prospective CC screening cohort, which showed that SOX1/PAX1 methylation exhibited improved risk stratification and disease predictive performance than cytology among HPV‐positive women. In addition, methylation‐based triage strategies could reduce the number of colposcopies for each detected CIN2+/3+ case and postpone follow‐up intervals. Thus, SOX1/PAX1 methylation might be a potentially objective triage biomarker in CC screening.
4.2. Interpretation
Chan et al. showed that positive combined SOX1/PAX1 methylation correlated with a 48.09% risk of CIN2+, while 13.18% for negative methylation [26]. Our results showed baseline CIN2+/3+ risks of 34.38/26.21% for positive SOX1/PAX1 methylation and 2.42/0.19% for negativity, suggesting combined negative results may be more reassuring for immediate CC screening in the current study. A prospective cohort of 1758 HPV‐positive women reported positive result of 6 gene methylation panel carried lower immediate CIN2+/3+ risks than our results in HPV‐positive women (12.7%/8.3% vs. 34.48%/26.21%), while negative methylation showed comparable CIN2+/3+ risks (1.3%/0.2% vs. 2.42%/0.19%). The similar results were also observed in HPV 16/18‐positive women and non‐16/18‐positive women [17]. More importantly, negative result of 6 gene methylation could lower 3‐year CIN2+ risk from 6.8% to 4.5% for HPV‐positive women, which suggested negative DNA methylation at baseline was associated with lower long‐term risk of developing CC [17]. Similar with previous data, our study revealed that negative SOX1/PAX1 methylation could lower 3‐year CIN2+/3+ risks from 15.98%/9.09% to 6.52%/1.86%. In this cohort, SOX1/PAX1 methylation showed strong risk stratification performance among HPV‐positive women, suggesting their potential as predictors of long‐term disease progression.
Currently, cytology and p16/Ki‐67 dual stain (DS) are both common triage methods for HPV‐positive women. The observational study based on the KPNC cohort reported that DS could further stratify HR12‐positive women with NILM cytology into high and low CIN3+ risk groups (16.8%/7.5% of CIN2+/3+ risks for DS+ and 6.6%/1.6% for DS−, respectively) [27]. Another study based on the IMPACT cohort showed 14.8%/3.4% of CIN2+/3+ risks for DS+ and 3.5%/1.0% for DS− among HR12‐positive women with NILM cytology [28]. Our results found that 8.22%/6.85% of CIN2+/3+ risks are for positive methylation, which should be referred to colposcopy according to ASCCP guidelines, and 0.99%/0.25% of CIN2+/3+ risks are for negative methylation, which were lower than the CIN2+/3+ risks reported for DS‐negative women in prior studies of similar populations. Thus, SOX1/PAX1 methylation might appear to have a potentially favourable or comparable performance to cytology‐ and DS‐based triage for hrHPV‐positive women. However, external validation and head‐to‐head comparisons with DS‐based triage strategies are needed.
Multiple studies have verified varied methylation panel could provide good CC screening efficacy in the triage of hrHPV‐positive women, including PCDHGB7 hypermethylation, S5 methylation classifier, and FAM19A4/mir124‐2 methylation, and with the sensitivities of 31.6%–96.6%/69.9%–100.00%, specificities of 47.1%–91.1%/56.6%–74.5% for detecting CIN2+/3+, respectively [17, 29, 30, 31, 32, 33]. Recently, Chan et al. evaluated the triage capability of combined PAX1/SOX1 methylation using samples from a prospective randomised controlled trial, reporting sensitivity of 75.21% (67.39–83.04) and specificity of 66.78% (61.32–72.24) in detecting CIN2+ [26]. Compared with previous studies of methylation detection, our results suggested that SOX1/PAX1 methylation‐based triage for hrHPV‐positive women offered similar sensitivity but higher specificity. Thus, SOX1/PAX1 methylation‐based triage method might be a complementary option to the existing approaches.
Our study also found that SOX1/PAX1 methylation significantly improved sensitivity, specificity, PPV, and NPV for identifying both immediate and 3‐year cumulative CIN3+ cases compared to cytology triage. However, the detection performance of CIN2+ was not significantly improved except the specificity and/or PPV. It might be attributed to the variable and unpredictable natural history of CIN2, which includes regression or further progression, while CIN3 is a precancerous lesion that has a higher risk of progression to CC compared with CIN2 [34].
Wentzensen et al. have found that DS triage for HPV‐positive women increased the specificity from 46.5% to 53.1% in the detection of CIN3+ compared with triage only for HR12‐positive women [35]. Similarly, our results showed methylation triage for all HPV‐positive women significantly increased the specificity for immediate and 3‐year cumulative CIN2+/3+ compared with triage only for HR12‐positive women. More interestingly, optimised methylation triage strategy for combined HPV‐positive and cytology results also further improves the screening performance and needs fewer colposcopies per detected CIN2+/3+ case compared with cytology triage. Thus, as an objective method, SOX1/PAX1 methylation showed triage performance that may support its further evaluation as an option for hrHPV‐positive women, alongside cytology and DS, which could significantly reduce unnecessary colposcopy referrals and clinic visits, thereby streamlining the management of CC screening.
4.3. Strengths and Limitations
Our study demonstrates several strengths. Firstly, the prospective 3‐year follow‐up design within an organised, multicenter screening program provides robust evidence for both immediate and long‐term risk assessment; secondly, we conducted comprehensive modelling of multiple triage strategies using the same cohort data, enabling direct comparison of different management approaches within an identical population context; finally, our focus on both immediate diagnostic accuracy and longer‐term cumulative risks aligns with contemporary risk‐based management principles, representing the first comprehensive evaluation of SOX1/PAX1 methylation triage with extended follow‐up in HPV‐positive women.
There were also several limitations. Firstly, one subset (Xinmi) was excluded due to unavailability of cytology residue, which might lead to the potential selection bias; secondly, as colposcopy was not performed in all women with HR12‐positive and NILM cytology, which might affect the evaluation of the methylation triage efficacy; thirdly, the small sample size of certain subgroup, such as CIN3+ events, limited the precision of estimates; finally, our comparisons among multiple triage strategies were conducted without multiple testing. As this study was intended to be exploratory in nature, and the generated hypotheses in the current study need further external validation.
5. Conclusions
Our cohort showed that SOX1/PAX1 methylation could improve risk stratification compared to cytology among HPV‐positive women, with high sensitivity and specificity for CIN2+/3+ detection in CC screening. This approach may help identify cervical lesions as many as possible while avoiding unnecessary colposcopy referrals. Additionally, methylation‐based triage strategies could potentially postpone follow‐up intervals beyond the standard 3‐year screening period, possibly even improving screening efficiency. However, external validation and head‐to‐head comparisons with other triage methods are needed before clinical implementation.
Author Contributions
Hongyu Xie, Xiao Li, Weiguo Lu, and Hui Wang conceived and designed the study. Hongyu Xie and Ansong Emmanuel analysed the data. Hongyu Xie, Xiao Li, and Junyan Li drafted the manuscript. Huiru Huang, Weiguo Lu, and Hui Wang revised the manuscript. Hongyu Xie, Ansong Emmanuel, Xiao Li, and Hui Wang collected the data and critically revised the manuscript. Huiru Huang and Weiguo Lu verified the underlying data. All authors agreed to be held accountable for all aspects of this work and approved the final version of the manuscript.
Funding
This study was provided by the National Key R&D Program of China (2021YFC2701204), Key Research and Development Program of Zhejiang Province, China (2023C03169 and 2020C03025), Zhejiang Traditional Chinese Medicine Multidisciplinary Innovation Team for Abnormal Uterine Bleeding.
Ethics Statement
This study was approved by the Ethics Committee of Women's Hospital, School of Medicine, Zhejiang University (approval number: IRB‐20240112‐R).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Table S1: The distribution of cytology and histology in strata of HPV genotypes and methylation.
Table S2: The risk of CIN2+/3+ in strata of HPV genotypes among HPV‐positive women.
Table S3: The Colposcopy/CIN and Visit/CIN were used over the 3‐year period of CC screening by different screening strategies.
Table S4: The management recommendation based on the risks of CIN2+/3+ in strata of HPV genotypes, cytology and methylation among HPV‐positive women.
Contributor Information
Hui Wang, Email: wang71hui@zju.edu.cn.
Weiguo Lu, Email: lbwg@zju.edu.cn.
Xiao Li, Email: 5198008@zju.edu.cn.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Castle P. E., Einstein M. H., and Sahasrabuddhe V. V., “Cervical Cancer Prevention and Control in Women Living With Human Immunodeficiency Virus,” CA: A Cancer Journal for Clinicians 71, no. 6 (2021): 505–526. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Fontham E. T. H., Wolf A. M. D., Church T. R., et al., “Cervical Cancer Screening for Individuals at Average Risk: 2020 Guideline Update From the American Cancer Society,” CA: A Cancer Journal for Clinicians 70, no. 5 (2020): 321–346. [DOI] [PubMed] [Google Scholar]
- 3. Perkins R. B., Guido R. S., Castle P. E., et al., “2019 ASCCP Risk‐Based Management Consensus Guidelines for Abnormal Cervical Cancer Screening Tests and Cancer Precursors,” Journal of Lower Genital Tract Disease 24, no. 2 (2020): 102–131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ronco G., Dillner J., Elfström K. M., et al., “Efficacy of HPV‐Based Screening for Prevention of Invasive Cervical Cancer: Follow‐Up of Four European Randomised Controlled Trials,” Lancet (London, England) 383, no. 9916 (2014): 524–532. [DOI] [PubMed] [Google Scholar]
- 5. Zhao Y., Bao H., Ma L., et al., “Real‐World Effectiveness of Primary Screening With High‐Risk Human Papillomavirus Testing in the Cervical Cancer Screening Programme in China: A Nationwide, Population‐Based Study,” BMC Medicine 19, no. 1 (2021): 164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Aarnio R., Wikström I., Gustavsson I., Gyllensten U., and Olovsson M., “Diagnostic Excision of the Cervix in Women Over 40 Years With Human Papilloma Virus Persistency and Normal Cytology,” European Journal of Obstetrics & Gynecology and Reproductive Biology: X 3, (2019): 100042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Mabeya H., Khozaim K., Liu T., et al., “Comparison of Conventional Cervical Cytology Versus Visual Inspection With Acetic Acid Among Human Immunodeficiency Virus‐Infected Women in Western Kenya,” Journal of Lower Genital Tract Disease 16, no. 2 (2012): 92–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Luckett R., Mogowa N., Li H. J., et al., “Performance of Two‐Stage Cervical Cancer Screening With Primary High‐Risk Human Papillomavirus Testing in Women Living With Human Immunodeficiency Virus,” Obstetrics and Gynecology 134, no. 4 (2019): 840–849. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Winer R. L., Hughes J. P., Feng Q., et al., “Early Natural History of Incident, Type‐Specific Human Papillomavirus Infections in Newly Sexually Active Young Women,” Cancer Epidemiology, Biomarkers & Prevention: A Publication of the American Association for Cancer Research, Cosponsored by the American Society of Preventive Oncology 20, no. 4 (2011): 699–707. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Wang Y., Mao Y., Wang C., et al., “RNA Methylation‐Related Genes of m6A, m5C, and m1A Predict Prognosis and Immunotherapy Response in Cervical Cancer,” Annals of Medicine 55, no. 1 (2023): 2190618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Fang C., Wang S. Y., Liou Y. L., Chen M. H., Ouyang W., and Duan K. M., “The Promising Role of PAX1 (Aliases: HUP48, OFC2) Gene Methylation in Cancer Screening,” Molecular Genetics & Genomic Medicine 7, no. 3 (2019): e506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Yang X., Chen Y., Li M., and Zhu W., “ERBB3 Methylation and Immune Infiltration in Tumor Microenvironment of Cervical Cancer,” Scientific Reports 12, no. 1 (2022): 8112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Chen Y., Cui Z., Xiao Z., et al., “PAX1 and SOX1 Methylation as an Initial Screening Method for Cervical Cancer: A Meta‐Analysis of Individual Studies in Asians,” Annals of Translational Medicine 4, no. 19 (2016): 365. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Kan Y. Y., Liou Y. L., Wang H. J., et al., “PAX1 Methylation as a Potential Biomarker for Cervical Cancer Screening,” International Journal of Gynecological Cancer: Official Journal of the International Gynecological Cancer Society 24, no. 5 (2014): 928–934. [DOI] [PubMed] [Google Scholar]
- 15. Zhang W., Wang H., Chen S., et al., “Reactivation of Methylation‐Silenced PAX1 Inhibits Cervical Cancer Proliferation and Migration via the WNT/TIMELESS Pathway,” Molecular Carcinogenesis 63, no. 7 (2024): 1349–1361. [DOI] [PubMed] [Google Scholar]
- 16. Lin Y. W., Tsao C. M., Yu P. N., Shih Y. L., Lin C. H., and Yan M. D., “SOX1 Suppresses Cell Growth and Invasion in Cervical Cancer,” Gynecologic Oncology 131, no. 1 (2013): 174–181. [DOI] [PubMed] [Google Scholar]
- 17. Zhang L., Zhao X., Hu S., et al., “Triage Performance and Predictive Value of the Human Gene Methylation Panel Among Women Positive on Self‐Collected HPV Test: Results From a Prospective Cohort Study,” International Journal of Cancer 151, no. 6 (2022): 878–887. [DOI] [PubMed] [Google Scholar]
- 18. De Strooper L. M. A., Verhoef V. M. J., Berkhof J., et al., “Validation of the FAM19A4/mir124‐2 DNA Methylation Test for Both Lavage‐ and Brush‐Based Self‐Samples to Detect Cervical (Pre)cancer in HPV‐Positive Women,” Gynecologic Oncology 141, no. 2 (2016): 341–347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Rogeri C. D., Silveira H. C. S., Causin R. L., et al., “Methylation of the Hsa‐miR‐124, SOX1, TERT, and LMX1A Genes as Biomarkers for Precursor Lesions in Cervical Cancer,” Gynecologic Oncology 150, no. 3 (2018): 545–551. [DOI] [PubMed] [Google Scholar]
- 20. Zhu P., Xiong J., Yuan D., et al., “ZNF671 Methylation Test in Cervical Scrapings for Cervical Intraepithelial Neoplasia Grade 3 and Cervical Cancer Detection,” Cell Reports Medicine 4, no. 8 (2023): 101143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Lai H. C., Ou Y. C., Chen T. C., et al., “PAX1/SOX1 DNA Methylation and Cervical Neoplasia Detection: A Taiwanese Gynecologic Oncology Group (TGOG) Study,” Cancer Medicine 3, no. 4 (2014): 1062–1074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Li X., Rao X., Wei M. J., Lu W. G., Xie X., and Wang X. Y., “Extended HPV Genotyping for Risk Assessment of Cervical Intraepithelial Neoplasia Grade 2/3 or Worse in a Cohort Study,” Journal of the National Comprehensive Cancer Network 20, no. 8 (2022): 906–914.e10. [DOI] [PubMed] [Google Scholar]
- 23. Solomon D., Davey D., Kurman R., et al., “The 2001 Bethesda System: Terminology for Reporting Results of Cervical Cytology,” JAMA 287, no. 16 (2002): 2114–2119. [DOI] [PubMed] [Google Scholar]
- 24. Egemen D., Perkins R. B., Clarke M. A., et al., “Risk‐Based Cervical Consensus Guidelines: Methods to Determine Management if Less Than 5 Years of Data Are Available,” Journal of Lower Genital Tract Disease 26, no. 3 (2022): 195–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Clarke M. A., Cheung L. C., Castle P. E., et al., “Five‐Year Risk of Cervical Precancer Following p16/Ki‐67 Dual‐Stain Triage of HPV‐Positive Women,” JAMA Oncology 5, no. 2 (2019): 181–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Chan K. K. L., Liu S. S., Lau L. S. K., et al., “PAX1/SOX1 DNA Methylation Versus Cytology and HPV16/18 Genotyping for the Triage of High‐Risk HPV‐Positive Women in Cervical Cancer Screening: Retrospective Analysis of Archival Samples,” BJOG: An International Journal of Obstetrics and Gynaecology 132, no. 2 (2025): 197–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Wentzensen N., Fetterman B., Castle P. E., et al., “p16/Ki‐67 Dual Stain Cytology for Detection of Cervical Precancer in HPV‐Positive Women,” Journal of the National Cancer Institute 107, no. 12 (2015): djv257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. T. C. Wright, Jr. , Stoler M. H., Ranger‐Moore J., et al., “Clinical Validation of p16/Ki‐67 Dual‐Stained Cytology Triage of HPV‐Positive Women: Results From the IMPACT Trial,” International Journal of Cancer 150, no. 3 (2022): 461–471. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Chang C. L., Ho S. C., Su Y. F., et al., “DNA Methylation Marker for the Triage of hrHPV Positive Women in Cervical Cancer Screening: Real‐World Evidence in Taiwan,” Gynecologic Oncology 161, no. 2 (2021): 429–435. [DOI] [PubMed] [Google Scholar]
- 30. De Strooper L. M. A., Berkhof J., Steenbergen R. D. M., et al., “Cervical Cancer Risk in HPV‐Positive Women After a Negative FAM19A4/mir124‐2 Methylation Test: A Post Hoc Analysis in the POBASCAM Trial With 14 Year Follow‐Up,” International Journal of Cancer 143, no. 6 (2018): 1541–1548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Adcock R., Nedjai B., Lorincz A. T., et al., “DNA Methylation Testing With S5 for Triage of High‐Risk HPV Positive Women,” International Journal of Cancer 151, no. 7 (2022): 993–1004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Verhoef L., Bleeker M. C. G., Polman N., et al., “Performance of DNA Methylation Analysis of ASCL1, LHX8, ST6GALNAC5, GHSR, ZIC1 and SST for the Triage of HPV‐Positive Women: Results From a Dutch Primary HPV‐Based Screening Cohort,” International Journal of Cancer 150, no. 3 (2022): 440–449. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Cao D., Yang Z., Dong S., et al., “PCDHGB7 Hypermethylation‐Based Cervical Cancer Methylation (CerMe) Detection for the Triage of High‐Risk Human Papillomavirus‐Positive Women: A Prospective Cohort Study,” BMC Medicine 22, no. 1 (2024): 55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Tainio K., Athanasiou A., Tikkinen K. A. O., et al., “Clinical Course of Untreated Cervical Intraepithelial Neoplasia Grade 2 Under Active Surveillance: Systematic Review and Meta‐Analysis,” BMJ (Clinical Research Ed) 360,(2018): k499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Wentzensen N., Clarke M. A., Bremer R., et al., “Clinical Evaluation of Human Papillomavirus Screening With p16/Ki‐67 Dual Stain Triage in a Large Organized Cervical Cancer Screening Program,” JAMA Internal Medicine 179, no. 7 (2019): 881–888. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: The distribution of cytology and histology in strata of HPV genotypes and methylation.
Table S2: The risk of CIN2+/3+ in strata of HPV genotypes among HPV‐positive women.
Table S3: The Colposcopy/CIN and Visit/CIN were used over the 3‐year period of CC screening by different screening strategies.
Table S4: The management recommendation based on the risks of CIN2+/3+ in strata of HPV genotypes, cytology and methylation among HPV‐positive women.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.